Wildfires are growing in size and strength. With each year, these fires seem to cause more damage, making headlines and taking a toll on communities. But with each fire, we also see a pattern: governments and traditional methods struggle to keep pace.
In other industries, private companies have made significant strides where governments were unable to meet the swift changes in tech, economy and nature.
I asked myself recently: Could the private sector, with its ability to absorb tech integration, be a scalable and rapid answer to our expanding wildfire (and moreso – disasters in general) problem?
Wildfires Growing Wild
Year to date in 2023, the NIFC has recorded 88 large wildfires and 40k fires in total across the USA, amounting to 2M of acres burnt. 20’000 personnel were dispatched in response.
Growing wilder: An increase of +45% of annual fires and a stunning +75% of acres burnt per year are observed over the last ten years.
NIFC. Bottom two numbers in blue are: #of fires and acres burnt across the USA
Where Traditional Methods Fall Short
Government firefighting teams are skilled. They are dedicated. But they face challenges. With limited budgets, they often do not use the most modern (tactical) tools. Finding talent is a growing issue (as it is everywhere). Semi-Public jobs aren’t necessarily going to be the winners of finding talent…
And as fires grow larger, these teams are spread thin. There’s a need for help, and a private firefighting force usin cutting-edge technology might have the answers?
Lessons from Space: The Rise of Private Enterprises
When we look at space exploration, we see a clear shift. Once dominated by government agencies like NASA, the arena has dramatically transformed thanks to private companies. Companies like SpaceX, Blue Origin, and Isar Aerospace have changed the game. They’ve brought fresh ideas, swift implementation, and innovative technologies. SpaceX, for instance, reduced the cost of launching payloads to space, turning the dream of commercial space travel into a tangible reality. Blue Origin focuses on building a future where millions of people live and work in space. Meanwhile, Isar Aerospace is putting Europe on the map with its satellite launch solutions. These companies, unburdened by the bureaucratic pace of government agencies, have made space exploration faster, cheaper, and more ambitious.
The Same Happens on the Battlefield
A similar transition is evident on the tactical battlefield. Traditionally, national military forces were the primary actors in conflict zones. But in recent years, private military corporations (PMCs) have carved a significant niche for themselves. These corporations offer specialized services, from security detail for diplomats to logistics and support in war-torn regions. Their agility and capacity to provide tailored solutions have made them invaluable in many situations. PMCs can operate without the same level of red tape that national military units face. This allows for faster decision-making and deployment. Critics argue that this flexibility can sometimes border on recklessness or undermine diplomatic efforts. However, there’s no denying that PMCs have transformed the tactical landscape, proving that private entities can impact areas traditionally overseen by state actors.
And Also in Previously Public Healthcare
The healthcare landscape is also experiencing a seismic shift due to private tech intervention. Traditional healthcare systems, often mired in bureaucracy and slow to adapt, are now being supplemented or even bypassed by agile tech companies. These companies are revolutionizing the sector, offering solutions ranging from AI-powered diagnostics to telemedicine platforms. For instance, tech giants like Apple and Google are entering the health space with apps and devices that monitor vital statistics, enabling early disease detection and personal health management. Startups are also flourishing, offering telehealth consultations, digital therapy sessions, and online pharmacy services. By doing so, they’re making healthcare more accessible and tailored to individual needs. While these advancements promise improved patient outcomes and cost-saving measures, they also underscore the trend: the boundary between the tech industry and healthcare is blurring, heralding a new era of privatized, tech-driven health solutions.
Apply This to Disaster Response: The Potential of Private Tech in Wildfires
As you can see, private tech startups have already transformed many sectors. They bring tech absorption. They have money to invest in research. And they can act fast. Here’s how they might help with wildfires:
Tech:
Wildfire Prediction: Use data analytics to predict potential wildfire breakout areas by analyzing factors like dryness, vegetation, wind patterns, and historical data.
AI-based Dynamic Pathfinding: Use gen-AI plus aerial imagery and rapidly advancing radar techniques such as SAR (to “see” through forest canopies for example) to map and label the ground and terrain conditions. Use it to find the best path in real-time.
Robotic Equipment: Use robotic solutions for situations deemed too dangerous for human intervention, such as building firebreaks or navigating highly volatile areas.
Real-time Monitoring: Implement a network of sensors and satellite imagery to detect and monitor wildfires in real-time, giving instant updates about the fire’s direction, intensity, and speed.
Integrated Communication: Ensure seamless communication among teams, HQ, and external partners (like local fire departments) using encrypted channels, ensuring a swift and coordinated response.
Augmented Reality (AR): Equip ground teams with AR devices (like smart helmets) that overlay real-time data about the fire, safe routes, water sources, and wind patterns.
Resource Allocation: Analyze available resources (equipment, personnel, and more) and dispatch them efficiently based on real-time needs and the scale of the fire.
Employment Experience:
Better Pay and Benefits: Provide significantly better compensation to veteran personnel leaving the public sector, as well as rookies wanting to join the job and deciding between public and private sector.
Advanced Equipment: Use cutting-edge technology to appeal to firefighters wish to be effective in the field.
Integrated Dynamic Pathfinding and response to operational threats in real-time.
More Safety? Apply technology to safety protocols, but also to prevent dangerous situations in the first place.
Operational Deployment:
Aerial Support: Advanced use of drones for both reconnaissance, active firefighting (e.g., water drops) and supply lifts. Drone fleets can be used to monitor fire progression and direct ground crews more effectively.
Rapid Response Units: Develop specialized teams ready to be airlifted or mobilized at a moment’s notice to the heart of the fire outbreak, equipped with advanced firefighting equipment and communication tools.
Training: Offer specialized training to firefighters, focusing on both the technical aspects of firefighting and physical endurance and tactics. This would combine firefighting best practices with some of Blackwater’s expertise in operating under extreme conditions.
Business Model:
Subscription: Local governments, state parks, and private landowners can subscribe to the service, ensuring rapid response in case of a wildfire on or near their property.
Fee for Service (essentially paid-per-response): Offer disaster response, but also data analysis, tech solutions, and training to other firefighting agencies or private entities worldwide.
Equipment Sales: Develop and sell specialized firefighting equipment and technologies to other parties.
Here are some issues you might face if founders want to go after this big opportunity:
Coordination with Local Entities: Ensuring seamless collaboration with local fire departments will be crucial.
Regulatory Hurdles: The deployment of a private firefighting force may face resistance or regulatory issues in many regions.
Unions: Especially in US, what is the implication to work with the IAFF and smaller unions? Do they have any leverage?
Culpability: Who is gets what blame when customer (state) and supplier (private corps) disagree on the effectiveness and timing of the disaster response?
Why Now, You Ask ?
Consider this:
Growing Frequency and Intensity of Wildfires: Climate change has significantly contributed to more extended fire seasons, drier conditions, and more intense wildfires. Regions previously untouched or minimally impacted by wildfires are now seeing increased activity. This escalating threat necessitates innovative solutions. In fact, the US has already called in the National Guard (i.e. Military) to help with responding to massive wildfires in 2020 and 2021.
Failed States → Private Companies: As previously seen in defense, space and healthcare, state-critical sectors that crumble under massive requirements when administered by state-actors get “fixed” by outsourcing narrowly scoped tasks/projects to mega-private corps who can work more agile, adopt technology, and attract talent.
Rapid Technological Advancements > Tech Absorption of Incumbent Firefighting Forces: Data analytics, AI, satellite imaging, robotics and drone technology have reached a level of sophistication that allows for accurate wildfire prediction, real-time monitoring, and effective firefighting strategies.
Limited Resources of Public Firefighting Entities: In many regions, public firefighting resources are stretched thin. As wildfires become larger and more frequent, public entities are struggling to respond effectively. A private corps could provide additional manpower and resources during peak times.
Public Awareness → Political Considerations: Public consciousness about wildfires, largely driven by high-profile fires and their coverage in the media, is at an all-time high. While the ability to fight them is deteriorating. Disaster responses from history can “make or break” a political candidate in an election year, see Katrina (2006) in the US or the Oder-Floodings (2002) in Germany. Natural disasters – and the ability to effectively respond to them – creates political exposure to incumbent administrators, who in turn might be very inclined to adopt “easy” private solutions to outsource the problem-solution and maintain exculpation.
I remain extremely interested in this opportunity. We (Foundamental) are affiliated with large forest owners. As investors in the real-world, packed with this understanding of forests, and our history of partnering with founders in hard-tech, 3D and robotics, we are quite intrigued by what this opportunity has to offer to build a vast company to improve earth.
I found it interesting that the number of M&A deals from January through July 2023 is almost on pace with the record year 2021 for the same timeframe.
But it‘s curious, because I had heard several complaints by M&A professionals on the buy-side in construction-tech how painful the M&A process has become for them this year.
It’s clear: Doing strategic acquisition deals can be a huge value unlock for acquirers and founders. But they come with their own set of headaches.
So I set out to speak with 15 M&A pros on the buy-side at construction companies, software strategics, private equity shops, and tech unicorns, all of them active in acquiring construction tech firms.
Some recurring themes popped up on their pains and desired solutions.
Limited access to the right targets, lots of noise
The biggest frustration repeatedly mentioned is getting access to solid targets, especially in the middle market. As one Head of M&A put it, “I’m out elephant hunting but all I find is shitty VC-backed deals.” Many strategics told me they feel like they see the same recycled deals from bankers over and over, not proprietary deal flow. They want pre-vetted targets curated specially for them.
Tied to this, I heard almost in every meeting that acquirers feel there is too much noise and just not enough quality companies in the market. With very colorful language, one long-time acquirer of construction software startups told me how there is too many low-quality venture-backed firms from the 2018-2021 vintage with high burn and nothing achieved, and this person felt products were too undifferentiated (“I have seen more than 10 construction project management software firms who pitched us to buy them, and their products were nothing to write home about”).
One PE had a tactic worth replicating: When they wanted access to the best firms in a specific space that they had a thesis for in a new geography, they just bought a small advisory firm that had served the space and its targets for a long time. Buy acqui-hiring that firm, they integrated the firm’s partners into the PE and gave their new team the job to speak to all targets they deemed high quality and build the deal access.
Inability to build relationships pre-sale
Strategics mentioned over and over again that they feel they are screwed if they first see a target when a banker or M&A broker starts a process. They expressly mentioned the requirement to have time to build relationships with owners and management. As another corp dev head said, “If we first see a deal from a banker, I know we’ll lose.” Early access is crucial. For example, experienced M&A professional said if strategics get a peek at deals 6 months before bankers auction them off, that allows time to build proprietary relationships with the founders.
One tech unicorn CxO who had been tasked with M&A opined that “The best start to an M&A down the line is to do business with each other. We always start with a pilot and then some business. It makes a deal so much more likely.”
Getting dragged into banker/broker-led processes, too much risk in too short time
Banker- or broker-ked force strategics to overpay, take on more risk, and close with less diligence. Make no mistake – the professionals on the buy-side know this. While prior to 2022 they couldn’t do much about it, this time is different. They scrutinize not just the target more – they scrutinize the process much more than two years ago.
But bankers and M&A brokers love running broad auctions. While the strategics and PE pros I spoke to crave proprietary, bilateral deals avoiding formal competitive bidding wars.
As one M&A leader said, “Anytime I told my boss a deal came from a banker, he said forget it and find a better deal.”
Another senior PE investor mentioned,”We have stopped doing secondary PE deals, where a [construction software] firm is offered to us by another PE. We hunt for proprietary stuff in this market.”
Diligencing non-core targets
This is specific to the strategics I spoke with. Big enterprises mentioned they often struggle diligencing and integrating acquisitions outside their core-core expertise. One M&A leader gave an example of his company having lots of competence in rolling up small business in their core business, but when it comes to a software that is just targeting their core business, their competence how to evaluate it drops to zero, making M&A there painful. Strategics seem to want help diligencing non-core targets.
Unrealistic expectations, misaligned with market
Strategics specifically felt that too many founders still have sky-high price hopes misaligned with strategic buyers’ realities. One professional specifically mentioned how they are baffled how many founders “haven’t gotten the memo that it’s 2023.” This person felt that founders’ bankers rarely position the deal well or explain why the strategic is the best acquirer, making higher valuation expectations even less likely.
One M&A leader in a large industrial who does a ton of M&A in construction-related spaces mentioned: “The best predictor if we end up doing that deal is whether in the first meeting the founder tells us why they think we will be the best owner. […] The reason is it makes us believe that they want this deal with us.”
What acquirers find helpful
Proprietary access to off-market targets
Early looks at companies not openly for sale seems to be seen as a huge value. It is difficult especially for strategics, as founders rightly maintain information asymmetry. Strategics prefer preemptive access before bankers run wide auctions. A 6+ months heads-up lets strategics build relationships with founders-owners directly, thus making deals more likely. One corp dev head said a monthly preview of soon-to-be-for-sale companies would be clutch.
Diligence help on businesses they haven’t dealt with historically
Strategics struggle diligencing non-core targets. One M&A-experienced General Counsel said this is where external help adds tons of value, while they need little assistance on core deals. Outside expertise who can diligence non-core targets provide big utility.
Insights on founders’ mindsets
Understanding what drives a founder beyond money helps strategics craft better fitting offers. Learning founders’ priorities early lets strategics creatively structure deals using non-financial incentives important to founders. For example, one PE investor had cases where deep insights into startup founders’ motivations enabled better deal resolutions for both sides. Another corporate M&A head told me how one founder really wanted their name to be shown in their software for eternity, and how knowing this early can help both sides focus on what matters to them.
Funding startups to buy competitors
One move that some (especially software) acquirers seem to love is to lead funding rounds in startups whose products they deem competitive, or hugely synergistics, thus gaining access to the startup’s governance. Backing startups as a buyer, de-risking deals by leading rounds, and supporting rollups provides big value (to some strategics). One corp dev leader emphasized this strategy as extremely useful.
Well guess what, founders and existing investors of the best startups know this tactic all too well. I’m personally not convinced that this works for the top startups and category leaders in a space, but it actually leads to adverse selection in hot spaces.
Take-aways for acquirers in construction-tech
Do more business with the startups and scale-ups you have on your target list, starting with pilots and first contracts. Become important to each other
Build partnerships with VCs who can give you the lay of the land, and curated access to the top founders, instead of shallow and fast-paced banker auctions. The good VCs won’t tell you everything, but they know which founders are beginning to prepare an exit and if these founders want to speak to the strategic
Reference-check founders with their investors to find out if a founder cares about more than money, eg. autonomy, impact, prestige etc. and discuss deal terms accordingly
The devil’s tactic (not endorsed !!!): Proactively fund and take board seats in your name in key startups to support executing a buy-and-build strategy. I personally made the experience this is a surefire way to get into adverse selection problems and actually NOT get access to the best founders and firms, but hey, you do you. It has been mentioned by one buyer, so for completeness I am listing it here. I just think it’s an error-ridden tactic executed by some well documented strategics
Take-aways for ConTech startup founders
Note that the following conclusions refer to the insights specifically from 15 interviews I mentioned at the top, and are not exhaustive.
More importantly, they apply mostly to construction-tech startups plateauing in growth and margins.
For those able to sustain very rapid growth (2-3x+ revenue growth yearly) while achieving best-in-class margins, not all of these principles apply and usually you will find it more useful towards a bigger liquidity event down the line.
But for most construction tech ventures with an eye to a trade sale exit in 2-3 years, these seem to work:
Research an acquirer’s strategy and craft a thesis on why you’re the perfect fit for their goals, and then say that out loud in the first meeting
Give 6+ months ability to preferred acquirers to work with you, ideally through business, but at least through meetings. Make them comfortable with you as the owner and operator
Share your financial and non-financial motivations early in the process and get creative structuring deals to optimize for autonomy, impact etc.
In my personal opinion, avoid banker-led / broker-led processes unless you are a rockstar firm who can be a no-brainer deal for a large group of acquirers. If you are not, based on the shared learnings above, you seem actually less likely to strike the right deal in a banker-led process
Also personal opinion: Absolutely DO NOT ACCEPT THE DEVIL’S TACTIC above. It kills your optionality and moves all leverage to that one potential acquirer
The best deals come from invested relationship-building and proprietary processes focused on specific acquirers’ needs. With the right moves, founders can find the ideal strategic partner or private equity buyer and unlock life-changing value for themselves.
We have all heard how Software-as-a-Service (SaaS) multiples have come down their peaks two years ago.
But not just that. Being in Architecture-Engineering-Construction (AEC) technology, I can’t count how often and as recently as two weeks ago, I get to see PR how awesome some software firms think they are themselves, and how some of them frame themselves as this and that and whatever.
I like to work with numbers and facts. They tell the true story. They don’t always tell the entire story, but they do tell the true story.
So I went out and looked at a few publicly listed AEC software firms and their numbers, and benchmarked it with other public B2B software firms to let the numbers tell me the answers:
What’s the status of SaaS multiples in AEC? What do public investors reward with higher valuation multiples right now?
Can AEC support high(er) valuations as other, “sexier” B2B sectors?
Is PR of some AEC software firms backed up by substance?
What should current early-stage software founders in construction and architecture take away from this for their startup journey?
Here’s what I found, in charts.
I’ll keep the text shorter and let the numbers speak first. I posted my conclusions at the bottom.
Note that all numbers I used are publicly available, published by the respective companies. No proprietary or otherwise private information is used. Numbers are rounded. Primary source I used is Yahoo Finance, and the companies’ annual reports where Yahoo Finance did not provide data.
Let’s benchmark 3 AEC software firms to 3 other software firms that make substantial portions of their revenue with other B2B sectors:
Growth first: Procore ahead, although on a much smaller scale (baseline effect?)
What about margins: Who is actually running a best-in-class SaaS business?
Driven by acquisition and retention efficiency: The one SaaS metric that never lies – revenue per employee. Let’s see
While overhead efficiency can reveal how fat converts into revenue (or doesn’t)
And ultimately all of it needs to generate cash: Judge yourself
Now that you saw the metrics: What SaaS valuation multiples are public investors paying currently?
Lazy overview: Benchmarked in one chart
My personal opinion
This post is not about any of these firms. Draw your own conclusions about any of them.
My interest lies in what software founders in construction and architecture-engineering can learn from it for their startups and scale-ups.
So let’s circle back to the questions I opened with above:
What’s the status of SaaS multiples in AEC?
In this sample set, 6.7x to 13.7x on annual revenue (not ARR – that should be lower in most cases due to a forward multiple)
and 40x to 113x on operating profit (for the profitable firms).
Early-stage AEC SaaS founders should plan with their long-term forward ARR multiples to be in that same range, around 9-11x.
Earlier rounds can allow higher multiples, but often not substantially due to the long-term multiple compression. In practice I am seeing up to 40x on ARR prior to Series-A, around 20-25x at Series-A if growth and unit economics are very good (otherwise close to 10-15x), and post Series-B close to the long-term multiple.
What do public investors reward with higher valuation multiples right now (mid 2023)?
Clearly: Margins and sales & marketing efficiency, plus cash extraction and capital efficiency.
Growth might not be a differentiated factor in this sample-set as much as you think – the fastest growing firm comes from the lowest revenue and what you see is a baseline effect, not necessarily faster expansion. If you express the growth in absolute terms:
Autodesk added $1.7B in annual revenue since 2019
Bentley $350M
Procore $430M
Adobe $8.5B
Salesforce $14B
Dropbox $650M
If you still are a believer that growth narratives at this scale still get priced in – for how much longer do you think. Do you expect multiple compression in these cases unless efficiency catches up? Be your own judge.
The cases in this set that warrant closer inspection to understand their valuations (and whether you think they are correct) are Salesforce, Bentley and perhaps Procore. Draw your own conclusions, as I will keep my thoughts to myself.
Can AEC support high(er) valuations as other, “sexier” B2B sectors?
Yup, AEC software clearly can keep up with other B2B SaaS sectors in terms of valuations.
Is PR of some AEC software firms backed up by substance?
I conclude that some firms totally deserve their hype. They are great businesses.
While others might live more from constant storytelling and sales & marketing efforts than fundamentally sound business ?
To understand the difference – run the numbers.
Everything else is just propaganda ?
What should current early-stage software founders in construction and architecture take away from this for their startup journey?
Nail your go-to market motion before you scale. Construction is notoriously complex.
If you believe in the network and referral effects in construction software (like I do), your product needs to foster that.
… and your numbers need to prove it. If your numbers, especially ARR / FTE and S&M intensity and CAC metrics don’t show it, don’t feed yourself a BS narrative that you are a platform and have an amazing network-backed sales motion. You would be fooling yourself. Until your numbers prove it, you don’t have it. Go back to two bullets above.
Amidst the towering skyscrapers and bustling streets of modern cities lies a hidden menace that threatens the very fabric of urban life – Urban Heat Islands (UHIs). These heat-trapping zones are far from the idyllic tropical islands we envision; instead, they are urban areas that experience significantly higher temperatures than their surrounding rural regions.
In other words:
Cities emit heat (people, cooling/heating buildings and devices, transportation fumes)
Construction materials used in many cities absorb and retain heat
We rarely used any materials or devices with heat-reducing properties
There you go: your personal Urban Heat Island
Data first
MIT Climate Portal says that UHI’s can on average be 5-7 °F warmer than the surrounding countryside.
Sounds like not much ?
Well, consider this:
1 degree is the difference between water being in its liquid phase and frozen phase
the earth’s climate is changing due to less than that difference
7 °F can be the difference between your body feeling just fatigued or having a sunstroke
UHIs exist because of the concrete and asphalt that dominate the cityscape (for good reasons, because they still are among the most scalable building materials known to mankind).
But these surfaces absorb and retain heat, creating microclimates within urban areas that can be several degrees Celsius higher than their rural counterparts. The soaring temperatures in UHIs have far-reaching consequences that impact various aspects of urban life.
Health issues
As the urban thermometer climbs, vulnerable populations face more and more risks of heat-related illnesses. Elderly people, younger folk, and persons with pre-existing health conditions are extra susceptible to heatstroke and dehydration. Moreover, the relentless heat can exacerbate air pollution, worsen respiratory issues, and pose a serious threat to the well-being of city residents.
Energy consumption death spiral
What’s the immediate reaction to urban heat? You got it. Air conditioning. In a desperate bid to beat the heat, residents in urban heat islands (have to) resort to artificial cooling in buildings, homes, and transportation.
Thing is: This surge in energy consumption strains power grids and exacerbates greenhouse gas emissions, amplifying the impacts of climate change. The vicious cycle of UHIs and increased energy demands place cities at the forefront of the battle against global warming.
UHI >> disrupted weather >> more wear and tear in cities
What’s more is that the man-made disruption of the natural thermal balance within UHIs seems to have adverse effects on weather patterns, which would exacerbate the urban heat spiral.
If we continue lacking sufficient green spaces in cities, our urban environments will struggle to counteract rising temperatures through natural cooling mechanisms. And that scorching reality of UHIs places additional stress on city infrastructure, accelerating deterioration and increasing maintenance costs.
Wanted: Cool-er materials
Cities will be forced to take proactive measures to tackle UHIs and create resilient urban environments. Implementing green infrastructure, such as rooftop gardens, urban forests, and green corridors, can help cool cities by providing shade and increasing natural cooling through evapotranspiration.
Moreover, promoting the use of cool-er materials for building surfaces, along with energy-efficient design and construction practices, can help mitigate the urban heat trap.
What’s more is that reducing the heat island effect can enhance outdoor comfort, promote physical activity, and foster a sense of community cohesion.
Let’s not forget: Better transportation
Local governments can play a crucial role in combating UHIs by investing in sustainable transportation infrastructure, promoting public transportation, and encouraging active commuting methods like walking and cycling. If you look for examples where it works, I suggest you start your search with the Netherlands.
Such measures not only reduce the heat island effect but also mitigate traffic-related emissions and improve air quality.
Venture-backable opportunities
The UHI space is neither easy nor obvious for the traditional venture-backed founding opportunity. Replacing physical infrastructure – as in pavings, materials of existing buildings, transportation infrastructure – does not operate on a zero-marginal cost model (faaaaaaaaaaaaar from it) and also not on the short timelines ventures like to operate in.
Practically, it has sales cycles measured in decades, not days.
Just think how some of the most modern cities are not built “in situ” but from scratch (case in point: Neom), for that same reason that replacing legacy physical infrastructure is just a first-principles nightmare.
That said, I do suspect there exist venture-backable opportunities.
Just one example to leave with you all here:
Provision and maintenance of repeatable plant-matter for cities. The tech-enabled aspect here is the city planning, extremely efficient supply chain to standardize, procure, remove waste and install the plant matter, as well as a supply chain (skilled labor!) play to maintain it. Think: tech-enabled landscaper on city scale.
Would love to hear from you which other venture-backable opportunities you are seeing in the Urban Heat Island arena.
Since the beginning of 2023, my Foundamental colleagues and I have seen a huge inflection of tech founders and startups applying Neural Radiance Fields (NeRFs) to built-world objects.
NeRFs are state-of-the-art AI networks to generate 3D representations of an object or scene from a partial set of 2D images. For example, they even get reflections correctly represented in 3D. In a dumbed-down way, you can think of NeRFs as taking a 3D scan with just a couple of 2D photos from different angles.
Basically, NeRFs killed photogrammetry. (I won’t be writing about the technology behind NeRF’s in this post. You can read more about NeRFs in more detail here, here and here.)
This post is about my hypotheses what opportunities NeRFs create, for example in the building-world.
It’s an early and evolving post as my understanding of the applications of NeRFs and the market evolve. So I might update the post over time.
10 theses on big opportunities that NeRFs create, as of July 2023
Starting with a first principle bedrock: Through NeRFs, hundred of billions of objects and scenes will get captured. Because NeRFs in itself are open-source and easily adoptable, the NeRF tech will not be defensible nor differentiated in itself. The sheer scale of capturing of hundreds of billions of assets, however, will create new frictions in the 3D tool stack, and thus massive new opportunities.
Here are 10 current theses I play around with thinking about where value can get created in the context of massive adoption of hundreds of billions of NeRF-based assets.
☝️
Quick definition: In the context of this thesis, by “asset” we refer to 3D models, scenes and textures alike.
Upstream
Billions of new assets: NeRFs will capture and encode hundreds of billions of real-world objects and scenes in a matter of a few years.
Capturing is commodity: The upstream capturing and encoding with NeRFs will be commodity fast, and not sufficiently differentiated.
NeRFs are going to add hundreds of billions of real-world objects to our 3D economy. What will it do to the 3D tool stack, and what opportunities arise for founders?
Asset creation
More everyday authoring: The legacy 3D tool stack requires thousands of hours of practice to achieve routine and excellence. As 3D becomes accessible to new generations and user groups, modern authoring software built on concurrency (unlocking multi-player collaboration) and versioning reduces the entry barriers for casual 3D creators, and improves the integrations and open-ness for professional creators.
Reconstruction-as-a-process: Parametric is a thing, but it follows pre-determined rules and requires significant customization to the model or the scene. To animate a NeRF’ed object, or to re-texture it, or to change key aspects (eg. a chair with 3 instead of 4 legs), the tool-stack will require pipelines to AI-trained engines that automatically can upgrade NeRF’ed objects into new motion, texture or parametric changes. This shall include automated animation using trained models. It’s to be determined if this might require an infrastructure/API type deployment to make use of very advanced 3D infrastructure for the parametric system.
Magic tool-suites: A modern authoring tool is best when it does NOT try to fulfill every specialty function. That’s where magic tool suites come in. They are suites of ultra specialized functions, features and algorithms, which are interoperable with the authoring tool of choice.
Mastery
Decreasing marginal value, unless >>> Segmenting and tagging of real-world items: The flood of 3D’d real-world content could decrease the marginal value of access to yet another 3D asset, as discovery becomes a problem with every asset added. That’s why segmenting and tagging need to be automated. These are trained models to classify 3D assets on various vectors. While easier on everyday items that are google-able, it becomes an ULTRA technical problem and opportunity on non-googleable asset training data, eg. industrial assets or special mechanical parts of an object.
A new Digital Twin Management / PIM stack (especially for reconstructed built-world objects): To organize the discovery and re-use of assets via AI, new 3D product information management (PIM) software will be required to store, share and re-use (reconstructed) assets easily.
Distribution
4D NeRF streaming: This is a thesis I am currently most intrigued by. A true moonshot. Imagine a sports event, or a frontline defense use case, where objects and scenes get captured and encoded as the objects and scenes move through time (4D) and not re-live, but live. And those objects and scenes get encoded in (near) real-time into a 4D experience (eg. via an AR such as Apple’s Vision Pro headset) to the user. This moonshot, if feasible, requires an infrastructure layer or service to receive, encode and send back real-time NeRFs moving through time into live experiences for events, defense and actually any remote real-world collaboration.
E-commerce deployment: While this is a thesis many folks seem to think about lot, and I find actually really boring for its crass consumerism. It means the deployment of products in their digital 3D representation into 3D shopping environments. While I get the market, it should be noted it will require significant hardware penetration before inflection.
Full-stack consumer suite: Think Instagram, but for NeRF’d content. Capture-edit-distribution via a mobile phone app + community. A very smart founder told me: “Our phone screens are sadly 2D”. Something to think about.
My 10 theses on one chart: What opportunities in the 3D tool stack NeRFs are creating in the next years
Unclear to me how it will play out in the market
AI animation – on what data: NeRF’d objects are static (today). Creating animated 3D content is insanely complex and a mix of science, craftsmanship and art. Animators deserve every penny they get paid. But there’s just not that many of them. Now, how do we animate content from a static NeRF to a dynamic model and scene with AI? It’s actually a data availability problem. My bet is you will never have as much animated content to train on than you will have static content. Thus, one scalable solution might be to train on deducing movement from static content. We’ll see.
NeRFs as shells: As NeRF’d content (today) will capture the surface of objects, but not its inner contents, it creates billions of shell models, not solid models. I have yet to figure out what this fact will do with its scaling potential in which segments.
Attaching dimensions to NeRF’ed objects: I have yet to see a robust approach that attaches dimensions (cm, mm, inch) to NeRF’d models. Can anyone point me to a robust approach?
Have thoughts on my thoughts? I’d love to hear yours. Email me at patric foundamental com.
My concern for and fascination with the magnitude of the skilled workforce crisis in Western markets is increasing every week. You can find some of my other research and thoughts I am having here, here and here.
This week I spent time on skilled bricklayers, and why we are in for a very rough ride. The bricklayer shortage in Europe and USA might have the highest gap between how many we need and how unattractive the job is (alongside jobs such as truck drivers).
If you’re like me, when you hear “bricks”, first thing you think of is red/clay bricks.
But the market and variations of bricks is much much wider in use.
By quality, you differentiate first-class to fourth-class bricks. Differences occur in smoothness of surface, sharpness of edges, strength, durability and consistent coloring.
By raw material, you find bricks made of clay, concrete, fly ash, sand lime, earth with relatively high aluminum content (”fire bricks”) and others.
Across those variations, the UK market for example uses 2.6 billion bricks every year. That’s billion with a B.
When using a few educated assumptions, you come to the following consumption breakdown in the UK:
With these assumptions, ca. 50% of bricks are used in residential and 50% in commercial projects. This will not be exact, but directionally a useful estimate.
That’s for the material (bricks).
Moving to the labor supply – bricklayers
The UK has 60’000 bricklayers currently in work. It needs 75’000 bricklayers to fulfil all building demand every year. A 20% gap, 15’000 bricklayers short in the UK alone. The latest World Economic Forum “Future of Jobs 2023” report uncovered that bricklayer is THE job reported short in the most European countries. Not electricians, not plumbers. Bricklayers.
World Economic Forum, 2023
On average, a bricklayer CAN lay 500 bricks per work day. Before you shout at me – yes a bricklayer can lay 700-900 bricks per day in a perfect environment for one hour in a bricklaying competition, but not over a day and certainly not over a year. On average, it’s around 500 per day that they CAN do.
However, we know from the factual numbers that the average bricks / day that a human bricklayer lays is actually only 170 bricks per day. We know this by dividing 2.6 billion bricks per year / 60’000 bricklayers in work = 43’333 bricks per year per bricklayer. Scaled to 250 work days, it gives 173 bricks per work day per bricklayer. The difference to the potential 500 per day is – overhead, travel time, work site prep, holidays, sick days.
In the UK market today, a junior bricklayer right after their apprenticeship will make around GBP 35’000 annual salary, and with a few years experience scale up to GBP 60’000 per year. Again, that’s today’s market. It’s expected that the same experienced bricklayer in 5-10 years will command around GBP 90’000 per year.
Their employer will charge for the labor (not the material) on average ca. GBP 450 per day. Again, can be much more if job is more complex or demands exceptional quality, but on average GBP 450 per day in invoiced revenue for the bricklaying work is a good average.
What bricklaying robots can do
One robot can realistically lay 1’000 bricks per day, or 250’000 per year on a regular 250 work-days. That is not science fiction, it is being done, for example by this startup for US and European markets. A robot will need some form of supervision, but the supervision scales to a degree and the same supervisor can oversee a fleet of robots laying eg. 5’000 bricks per day.
By comparison, the human bricklayer lays 43’333 bricks per year.
Now, let’s be fair here and compare apples to apples. The robot doesn’t have to worry about overhead such as selling projects or writing invoices. That work remains with the human. We know from anecdotal evidence that this part of the overhead is ca. 50% of the work hours. Therefore, 43’333 / 50% = 86’667 bricks per year that a human could lay if they were to avoid the overhead the same way the robot does.
Comparing the 250k bricks for a bricklaying robot to the 87k bricks for a human bricklayer, that means one bricklayer robot can equal 2.9 human bricklayers.
In today’s salary terms, that’s a cost efficiency of GBP 170k per year. In future salaries, that cost saving will be closer to GBP 260k.
Even better: More revenue potential
The other side of the same coin is what the bricklaying company invoices to their customers.
The invoiced revenue is 0.9 GBP per brick laid.
Let’s forget that a human bricklayer has overhead, and let’s assume their employer could invoice 250 work days @ 500 bricks per day. Which is impossible with all the admin, travel, work site prep. But let’s say they can get there. That’s an invoiceable revenue of GBP 113k per year. If they pay an experienced bricklayer the GBP 60k p.a. today, and have 30% additional salary cost incl. fuel and car (GBP 78k total), that’s a 31% gross margin on the invoiced human labor revenue.
One robot could invoice GBP 225k in bricklaying revenue (without material) per year. Let’s say it’s leasing/depreciation and maintenance costs the same per year as the human bricklayer (it doesn’t, but let’s go with it). That’s a 65% gross margin on the invoiced labor revenue from the bricklayer robot. In total, it’s incremental revenue of GBP +113K per year that the bricklaying firm will make.
And the best thing is: the robot doesn’t want a salary increase, no sick time, no holiday, and no threat of leaving to join a local competitor or start their own business to compete with you…
Bricklaying robots will have massively positive impact on housing
I’ll keep this one short, check my numbers below:
This means: By deploying 5’200 bricklaying robots in the UK (15’000 bricklayers short / 2.9 bricklayer equivalent of one robot), we can close the entire 15k bricklayer gap. By closing that gap, the 5’200 robots will help build 160’000+ incremental buildings across single-family homes (SFH), multi-family homes (MFH) and commercial buildings every year.
That’s massive impact on the economy.
I used UK numbers as an exemplary market because sources were so easily available. The numbers will be directionally similar for US markets that use bricks, France or Germany.
I’ve been silent on here for 3 weeks since my last post.
Among other AEC-Techies and some legends in the community – Alain Waha of Buro Happold, David Basulto of Archdaily, and various Con-Tech founders – together we kept busy initiating a massive open-source project for founders in AEC-Tech.
The Wallhack launched last Friday, and is something I am convinced will be a massive resource for founders in construction, architecture and engineering technology.
The Wallhack is the most objective and curated database of venture capital investors in our space. It focuses on curating VC firms who are actively investing in construction, architecture, engineering, design, skilled labor, blue collar, and supply chain technology for the building world.
The project’s primary mission is to give founders and all other folks interested in the AEC-Tech and Construction-Tech investor landscape a fast and simple tool to research long-standing investors with steady capital in and around AEC-Tech. It aims to maximize the use of objective, curated information, and decrease “fund propaganda” and “PR narratives” (which we know VC’s are not necessarily ashamed of using ;)). So let’s get rid of that propaganda and let the objective facts speak. As a byproduct of all this, The Wallhack also allows investors with relevance for AEC-Tech to provide a useful overview of their investment firm’s affinity, activity and record.
As a result, founders can use The Wallhack to find their dream investors. Simple as that.
Fundraising in construction-tech has a lot of nuance, and successful startups typically bring together different investor types. The Wallhack initiative is making that a reality for more investors and more entrepreneurs.
This becomes a reality because The Wallhack does a few things right in my opinion:
It contains more than 1’000 VC investors who have actively deployed capital in our sector. Not every VC qualifies for the list though. The Wallhack has a quite specific set of criteria for an investor to be featured.
Only objective, curated data from the world’s most reputed VC data pools is used. Investors do not self-curate their numbers. (VC’s can still claim their profile and update their core info, such as URL, team members, content to be featured etc. – but not the quantitative data).
Data is updated weekly across more than 20’000 data points. That is possible because data pipes are automated.
Truly open-source. No monetization, no paywalls, no ads, no tracking.
Something important to all of us initiators was that it becomes a community-driven project that pierces through walls of opacity (hence the nerdy gaming name). Initiated by a bunch of folks who have spent many years in AEC-Tech and Construction-Tech and who all believe in the power of open-source and objective data. We all have day jobs that are somehow embedded in the AEC sectors. The Wallhack is a project we initiated on our evenings and weekends.
I am looking forward to helping the project expand its feature set and scope, and see hopefully thousands of founder teams in our space benefit from the massive resource that The Wallhack is.
I think The Wallhack is off to becoming the default resource for VC/investor research in Construction-Tech and AEC-Tech. Golden times for founders!
Ever seen these lists or maps of “these are the top 100 startups to watch” ? I know, I know – who hasn’t.
Well – I wanted to backtest how correct and good those VCs or journalists who publish these “best-of” lists or “market maps” are at curating ConTech ventures.
Being a full-time early-stage investor in ConTech as part of the global VC in the sector (Foundamental) – where we speak with several thousands of founders in our sector every year – I’ve had more and more the hunch that “best-of” lists and “market maps” often showed a picture distorted from the sector’s venture reality, and were not very good at highlighting the eventual long-term winners and category creators.
Outside-in PR propaganda ≠ Actual inside-in performance? This analysis was to backtest the “curation quality” and “prediction performance” of all the “best-of” lists and “market maps” created by other people by using objective data. See for yourself…
Methodology
1️⃣
Researched 20+ “best-of” lists and “market maps” for ConTech by VCs and journalists
* Google in incognito mode first; then duplicated the approach with DuckDuckGo
* only lists/maps no later than December 2021 (to allow for backtesting) and
no earlier than 2016 (before 2016, ConTech practically didn’t exist)
* boolean search used:
(”top 50” OR “top 100” OR “top 25” OR “top 30” OR “market map” OR "landscape") AND (”contech” OR “constructiontech” OR “construction-tech” OR “construction tech” OR “construtech” OR “in construction” OR “disrupting construction”) AND (”ventures” OR “startups”)
* searched through ALL search results; text and image results
* excluded PropTech-only firms (of which there were few, eg. Knotel and Juniper Square)
* did not use single-country lists except for USA
2️⃣
Extracted all 530+ mentioned ConTech firms, added objective data-points, and amended the list by the most-funded firms not on anyone’s “best-of” list or “market map”
* cross-referenced each firm with the lists/maps they were mentioned on
* added ConTech firms that have raised $30M+ but were missing on the sampled
lists/maps
* added data-points for each firm:
# of mentions, $ raised total, $ raised before mention, $ raised after mention, $ raised in last 18 months, status
* sources used for funding data: Tracxn first; if Tracxn shows no funding then double-checked with Crunchbase
3️⃣
Ran multiple analyses
* are pundits early or late?
* do they make or miss winners?
* do they promote substance or hype firms?
* if they were a fund: would they earn or burn money?
The sampled ConstructionTech “best-of” lists and “market maps” from 2016-2021
[open-source]
Sampled ConstructionTech “best-of” lists and “market maps” from 2016-2021
The 530+ ConTech ventures mentioned in the sampled "best of” lists and “market maps” from 2016-2021 [open-source]
⛔
Remember: What this is NOT
* first of all: it’s NOT a Foundamental list, also NOT my list!
* It’s NOT an endorsement or curation, and not a measure of quality or long-term success of any firm; remember that this is just a meta-collection of all the sampled “best-of” lists and “market maps” that other VCs and journalists created in the first place (and to demonstrate the advantages and disadvantages of such lists/maps is exactly the point of this exercise – so if you feel this list sucks, that’s my point !)
* It does NOT give an accurate global representation due to a huge bias in the sampled lists/maps; due to the strong bias represented by the pundits’ “best-of” lists and “market maps” towards US (and some European) firms, it contains very few Asian, African and Latin American ConTech start-ups
* It’s NOT complete; in fact we at Foundamental have spoken with over 10’000 ConTech-related founders over the last five years, which gives you an idea just how short “best-of” lists and “market maps” are
* It can NOT show always the correct funding data; we decided to use only funding data available on Tracxn and Crunchbase, even if we know internally that the public data is incorrect; we choose not to disclose non-public info
Sources for funding data: Tracxn, Crunchbase per June 1, 2023; no Foundamental data used
1/ Early or late to the party?: % of total capital raised before
and after mention
🤔
Reading: Looks like they are very late to the parties.
2/ Making or missing winners?: Total capital raised; firms
mentioned 3 times or more vs. firms least-mentioned but raised $20M+
🤔
Reading: They miss a lot more successful firms – too many.
3/ Substance or hype?: % of total capital raised in last 18 months;
firms mentioned 3 times or more vs. firms least-mentioned but raised $20M+
🤔
Reading: This one’s a killer. They seem to promote firms that were much lesser able to withstand the tough funding environment of the last 18 months – by far.
PR > substance for these pundits.
4/ Earning or burning?: “If these lists were a fund that invested
at time of list”: Estimate of hypothetical fund return
💡
Assumptions used:
$1M investment at time of mention per firm. 40% Dilution happened before, 20% dilution after. 48 months holding period. For firms that had 0 mentions, valuation at time of investment assumed as $30M (which is a point in time when people who know the sector should have seen every firm).
🤔
Reading: Not much interpretation needed. The data says it all: “best-of” lists and “market maps” are sub-par at picking ventures.
The reason is simple: The “best-of” lists and “market maps” have missed a large number of large category creators in ConTech consistently. The lists/maps contain a substantial bias towards (i) rather picking any kind of US firm over terrific non-US ConTech firms and (ii) picking many more charismatic, PR-ready founders with high-PR-value investors than they pick less charismatic founders who are excellent operators (because they don’t KNOW that they are excellent operators…)
These lists/maps also clearly seem to be sub-par at picking eventual winners before Series-A’s, which is why you won’t find many firms created since 2020 whom you might think highly of and feel like they are missing here. That’s part of the whole point…
Findings and call-to-action to ConTech founders
ConTech “best-of” lists and “market maps” can waste your time: The data is clear – many pundits are late, bad pickers, and promote hype instead of substance. Reasons are also clear – lack of insight, no proprietary view about operator performance, and hidden incentives.
Adverse selection, anyone?: Alright, get this – out of our top 15 financially-performing companies, only one (1) was ever on a list or market map in its first four years after founding. Let that sink in. Of course correlation is not causation – I am not saying being on a list or map will make your ConTech firm average. But what I clearly see is the correlation between being heads down and under the radar while you build a product and a business with your customers, and your long-term substantial performance in Construction-Tech. Hype does not sell, PR does not sell – only product and substance does, to the no-BS world of construction and architecture clients.
The company you keep: You likely don’t know the past and performance of your peers you get lumped together with on a list or market map. Do they do well? Do they have a good reputation behind the scene? Do they maybe have bodies in their closet? You likely don’t know, and that’s a risk for you. Because your customers and investors might know – they tend to have talked to and looked at many more folks in the space. By accepting to be on a list or market map, you might unknowingly associate yourself with company you don’t want to be mentioned in, and your customers or investors might take notice of such associations. This happens all the time.
I could keep going, but this post already got out of hand.
Let me just close with this call-to-action to you all ConTech founders:
The only stakeholders that matter in your early days are your customers and your talent. If you accept PR for your firm – do it only if it benefits you with your customers or with talent.
Do not accept PR about anything else, especially not investors. It’s an investor’s job to find you. We literally do nothing else the entire day. If this is the reason you accept PR, you are literally wasting your scarce time.
I am a huge fan of staying under the radar. I have seen first hand that ConTech firms can get dinged by old-school construction clients for being perceived as non-serious tech firms, while others who presented themselves as serious firms with no tech PR footprint at all grew quite fast. These are anecdotes, of course – but this general direction is solid advice.
For the last week I was working with a founder team in the construction cash flow space. It triggered me to look deeper into the payables, receivables and cash disbursement workflows in the construction industry in the US and UK markets.
As for material supply invoices, I was unable to find a good source. So I’ll have to run with a total guesstimate here – let’s say each sub-contractor makes 20 distinct materials purchase orders (POs), and each general contractor makes 200 materials POs. Let’s further assume each PO leads to 2 partial incoming invoices from materials suppliers. That gives 22x20x2 + 200×2 = 1’280 incoming materials invoices for an average project. Let’s round it to 1’300.
Now, let’s go to the contracting invoices themselves. We have an educated guess that each sub-contractors works for ca. 3 months. That makes 4.4 sub-contractor equivalents over 15 months project (22 / 15 x 3 = 4.4)
Let’s assume that the subs as well as the general contractor write outgoing invoices (or applications for payment, for that matter) twice per month per project. That gives (4.4 + 1) x 2 x 15 = 162 outgoing invoices upstream. Let’s round it to 200.
So overall, on an average 15 month project, we could be looking at 1’500+ invoices and money transfers.
Now: because each invoice creates a payable and a receivable, and requires them to be closed after cash is received we actually look at 6’000+ book entries per project.
Now let’s extrapolate to the number of construction projects:
The US makes up ca. one seventh of the global construction market.
Which means, there could be 7M residential construction projects worldwide per year.
Now, historically, residential is ca. 33% of all construction spending. The other 67% are performed across commercial, industrial and infrastructure projects. Because those projects carry significantly larger sizes, there should be proportionally less projects in those segments.
Let’s assume non-residential projects carry 10x larger spending per project. That means there could be 7M / 10 x 2 = 1.4M non-residential construction projects per year globally. Again, let’s round down to 1M.
In total, that would give 8M construction projects worldwide p.a.. (Broken down to the US, for example, that would be ca. 1.2M projects per year – sounds directionally right)
The 1’500 invoices and 6’000 book entries mentioned above happen over a 15 months timeline. We need to scale them down to 12 months:
Invoices:
1’500 / 15 x 12 = 1’200 per year per project
x 8M projects worldwide
= 9.6 billion invoices p.a. worldwide in construction (let’s round up to 10 billion)
The truth is likely somewhere in-between these 10B and the 55B.
4’800 book entries per year per project
= ca. 40 billion book entries p.a. globally in construction (and could be as much as 200B)
That’s a lot of cash flows and paper. Here’s a few big issues with this many invoices and book entries:
Massive working capital. Because the various construction stakeholders are traditionally financed from within the same supply chain, one party has to front cash to pay the other party. Incoming and outgoing payments don’t happen on the same day. The difference is the working capital required for construction.
Developers are at constant risk of a general contractor default. The more sequential invoices from sub to GC to developer, the bigger that risk.
Ugly accounting and treasury overhead
Error prone
We know that construction is a $10 trillion global market. Let’s say on average, there’s 30 days of working capital in the construction system. That would amount to $10T / 360 x 30 = $830B of construction working capital.
To cross-check whether this number sounds directionally right: that would be ca.:
$830B x 33% / 7M = $40K working capital per residential project
$830B x 67% / 1M = $550K working capital for a non-residential project
While these numbers won’t be accurate, they are in the right ballparks.
Now imagine you could eliminate one pass-through party in the construction invoice chain, and find a way to pay suppliers more directly. Thus you might be able to reduce the invoices and wires by as much as 50%, which would reduce the working capital by proportionate 50%, or as much as $400B globally, $60B in the US or $20K+ on an average residential project.
To summarize:
10B – 55B construction invoices (and cash transfers) per year
40B – 200B accounting book entries p.a.
$830B working capital
$40K working capital on a resi project and $550K on non-resi
$60B of working capital in US construction could be saved if someone could eliminate one pass-through party in the invoice chain
What I find interesting in this context is: Who bears the burden of these numbers? While the usual answer is the sub-contractor, I actually begin to think the long tail of developers bear the biggest pains. And their pains from these numbers are more painful the…
more different GCs they have to work with on their projects
smaller the GCs they work with
more volume and no. of projects per year they develop
That’s why I have become enamored with solutions fixing invoices and wires in the cash flow chain starting with real estate developers downstream to sub-contractors.
Here, I am continuing my thesis. Writing about the vectors through which incumbents are coming under attack, and their clients can unshackle.
I have noticed how some investors prefer to think their software thesis in terms of features. I don’t find that very helpful. Features are just a result of some larger macro shifts.
That’s why we dont think of our 3D design thesis in terms of features. We think in terms of market structure and mega shifts.
Let me begin my thesis by using a unique metaphor:
Imagine your body needs chronic treatment. Let’s say Diabetes. And folks tell you that you need insulin to treat it and function well.
There’s a dealer (not saying who) who has the reputation for always having insulin. You get it from him. But he sold you insulin at a teaser rate.
And he’s mixing some nasty stuff into your insulin which makes you addicted and dependent. You find it very hard to go to another dealer, because this guy’s insulin is just really hard to switch away from.
In the foreseeable future, his teaser rate will run out.
Now, the dealer comes to you and says “hey, this other drug goes super nicely with the insulin i’m already giving you, it’s great for your body”.
But really, what it does it creates even more chemical dependency of your body so that you can never get out of his insulin.
And you look at your wallet and you think: “hold on, when my teaser rate runs out, all the money in wallet will go to my one dealer. WTF”
“Now he wants to make me even more addicted?”
“No way; i have to get off his $hit, get my life in order and find a less abusive way to get my insulin. And while I’m rethinking it all, I should find an insulin that is not mixed with some of his nasty addictive stuff !”
“Oh, if I am making that huge life change anyway, I should also never again be dependent on a single dealer.”
Now, at this point in your realization, imagine there’s someone who tells you they actually got away from the dealer and since then built their own insulin stack. They tell you they began sourcing the different chemical components from trusted sellers and brewed their own insulin inhouse into the mix that is BEST FOR THEIR body.
That would be perfect, wouldn’t it?
You’d like to do that. You’ll probably need good support from others who are on the same journey as you. But because the vendor has left so many clients unhappy and addicted, you can easily find a huge group of people who have the same issue as you and you can consult each other and team up so you can all brew our own mix and become our own producers.
Within this metaphor, we find five interesting components:
the need for a great product to fulfill life-critical functions (”control my blood sugar”)
the product stack
the vendor’s incentives
the pricing model
the community
Before double-clicking on them: I find it noteworthy that most “pundits” or investors out there think about the feature set, when they think about disrupting the dealer in the above metaphor. And it is a key component – if I can’t offer a decent substitute product, I can’t compete. But: the key to change in this metaphor also lies in working with the system’s incentives, pricing, structure and community.
Let’s bring the metaphor into 3D design and work through the components one by one:
Attack vector 1: Life-critical functions
The drug (insulin) in the metaphor can be equivalent to many tools used in the 3D design stack, but probably the most critical one in the stack of 3D designers and architects is their central CAD/BIM authoring tool. Without which the users couldn’t fulfill their core functions.
The attack vector here is not a more feature rich authoring tool (i.e. an insulin that consists of more chemical components, for that matter). The required feature set is sophisticated, but very clear.
Rather, the attack vector in terms of enhancing the life-critical functions that I see is:
Concurrency: The ability to manipulate and process elements in the same model at the same time.
atomic distributed version control aka git: The ability to store the state of every atomic object of the model at any point in time (and not just time snapshots !) from n-number of users and thus give a complete historical trackability of edits, branches and commits to the 3D model.
n:n multi-player: The ability to have n-number of users collaborate in n-number of distributed software from different vendors, without having to have one central vendor (or server) participate (which would not be n:n).
AI-ready streaming of my proprietary data: The ability to normalize, ringfence and distribute a firm’s or team’s model and model history, so that the team can build their own large language model based on their ringfenced but huge amounts of proprietary data without having to share their data outside their team.
Unfortunately for decades-old legacy authoring tools (the insulin), the above functionalities cannot be easily added on top of an age-old stack. They form a very different core stack around which to build an authoring tool from the ground up, starting with the database and data query architecture. Any legacy authoring tool will have a high development effort to switch to a modern stack that supports the above functions, plus the effort to convince and retrain their users over time.
And that is the attack vector here. Modern software for the 3D design stack built on modern technologies that allow concurrency, git, n:n and AI-ready streaming, offering not a 1:1 substitute but a drastically improved alternative to the drug fix currently on offer.
Which brings me to the entire stack.
Attack vector 2: Product stack
The dealer in the metaphor benefits from the fact that his client cannot home-brew a superior or otherwise better or cheaper insulin himself.
The same thing happened for decades in 3D design. A vendor would offer the most complete authoring software or the most sophisticated specialty tool. A client could never rebuild the same software in house. But worse – a combined stack of tools was not available that – when stacked together – would be a superior solution.
That’s what’s rapidly changing now. Architecture firms and design clients around the world are in the process of re-evaluating their IT spend. What they realize is that that one vendor will soon eat up their entire IT budget. We are talking 2025-2027 here. So soon.
While the largest clients are developing some lighter parts of their stack inhouse (eg. Testfit tools), the most sophisticated parts of the 3D design stack are just not feasible to be built by AE firms or design firms. Inhouse is not an option for the most sophisticated parts of the software stack.
Fortunately for them, these clients begin to actively go into the market, and look for new vendors that fulfil parts of the “dealer’s” software. These clients are willing to stack different softwares together, which do their part of the stack more excellently that the “we-try-to-do-it-all-in-one” software from the untrustworthy vendor.
That’s the attack vector that the dealer does not have an answer to. AE and design firms re-stacking their toolset with software that is individually superior in specific functions, and thus re-building the entire stack of their firm one by one.
Now, the dealer has no reason to participate in this, because: incentives…
Attack vector 3: Vendor’s incentives
In the metaphor, the vendor has the incentive not to heal the patient, but to keep them addicted. Being addicted means coming back for more, eternally. Every time you come back, the vendor needs to sell the patient more per dose, because doses hit resistance. A vicious cycle.
A very similar dynamic plays out when a vendor has maximum market power and market share in a 3D design software product that is ultra core to the user’s company (eg. an architecture firm) and is very complex to learn. The more the designer uses the product, the more the designer stores their data in the vendors software, the higher the addiction.
The vendor then has the incentive to keep adding more and more addictive features, and never let their clients come out from under.
Fortunately for the architecture and 3D clients, though, this is exactly the attack vector here:
Because clients are aware of these incentives (it’s an open secret and visible to the naked eye), the vendor has no credibility of ever meaning well for the clients. Even if the vendor has a change of heart – the hurt is already done. The only reason his clients stay with him is because there is no superior alternative. He abuses them. The moment the vendor faces true superior competition, his reputation is so hurt that whatever he claims he wants to do right by his clients – his clients know that he abused them.
Therefore, any alternative vendor (ventures, for that matter) offering superior substitutes can credibly claim to mean much more well for the vendor’s clients and structure their product and pricing model in a way that demonstrates credibly a good-natured alignment to the client’s benefits (and not to the dealer’s). All you have to do is to structure your product and license in any way that tilts to the benefit of the client, eg. in terms of length of commitment, service, flexibility in using the software, and pricing model. And the existing vendor has no way to defend against this attack vector because he created a systematic issue for himself.
Which brings me to pricing.
Attack vector 4: Pricing
The vendor/dealer in the metaphor is used to a steady and constantly expanding revenue stream from his clients. It will be very hard to change the pricing in a way that tilts in the favour of clients, because (a) “why would I” and (b) the vendor would take home less money where everyone has gotten used to that steady stream and pile of cash. How does the vendor explain that at home?
He doesn’t, or at least not without major discussions and convincing.
That’s the attack vector here. New entrants being able to offer pricing models much more aligned with client outcomes.
Instead of prohibitive seat-based licenses for 3D software, and a hard crackdown on clients who “over-use” according to the vendor – new entrants can offer pricing models that allow them very fast growth and allow architecture and 3D clients much more flexibility and independence. For example:
Pay for volume tiers, don’t pay for unexpected usage spikes
Pay for unlimited use, don’t pay for over-use
Pay for higher-tier features, don’t pay when features not needed
Pay for service level access (especially great in combination with open-source software), don’t pay for custom development and support
Attack vector 5: Community
All clients source their product from the same vendor. That’s why they share the same pain, and they know they share it.
The moment a few of them start building their own product stack, and ask around who can help with sourcing the parts needed, collaborating together to re-engineer a solution – the movement begins. Because the pain is so high and the trust in the vendor is low.
The vendor’s incentives, of course, are to avoid that movement. Control the communication, make sure they don’t speak with each other and make sure that no collaboration happens outside the control of the vendor.
That’s the final attack vector in my list. Activating a nascent yet vibrant community, and unleashing a movement that is brewing under the surface. Arm them to re-stack their toolkit.
3D design has one of the most amazing technical communities out there. Made up from game developers to aerospace engineers to architects – they all build in 3D, and many of them code. They organize online, they exchange ideas and learnings.
And new entrants – especially those that have a credible open-source strategy – are the ones that will be in the position to channel the power of community through this attack vector on the dealer.
My thesis is that through these 5 five attack vectors, 3D design clients can unshackle themselves from the dealer in the metaphor.
So much, for now.
Coming next: In Part 3, I plan to write how I see the attack vectors currently form in the shape of specific company examples – which I deem most exciting to be dealer killers.
Germany has a stock of 20 million residential buildings, plus millions of commercial buildings.
While the initiative has its merits, and is commendable from a carbon reduction perspective – it leaves many building owners in Germany (B2C and B2B) with a lot of uncertainty what it might mean for them.
On top of which, many commercial owners of buildings are already dealing with complexity from the EU’s taxonomy to reduce their buildings carbon footprint.
This uncertainty for building owners is best met with facts about their building’s real substance and thermal footprint.
That’s exactly what building energy audits do, and why they are growing like crazy in Germany (and Europe).
Think of the building energy upgrade cycle in terms of 3 eternally-repeating phases:
Assessing, which consists of capturing the building substance, running complex normed calculations, prioritizing upgrades, and designing them, and also benchmarking the building on a granular level
Installing, from procuring material and labor to carrying out the upgrade project and assuming the risk for the project
Operating, incl. capturing the building’s power and thermal footprint, monitoring variable prices and making operational adjustments
The installation phase will be fixed soon by an abundance of players
Today, the installation phase is severely bottlenecked by a lack of tens of thousands of skilled workforce to carry out the upgrades, such as retrofitting insulation, windows, heat pumps or rooftop solar. The good news in this phase is, though, that a bunch of players have recognized this demand overhang opportunity. Lots of players have raised capital to build asset financing companies for solar (eg. Enpal), for gas boilers and more recently heat pumps (eg. Thermondo), or to bring efficiency to retrofit insulation (eg. Varm – highly recommended to put on your watchlist) or even for generic installation workforce serving the above mentioned players B2B (eg. Installion). Also, OEMs of boilers and heat pumps try to play a big role in this phase.
So, fortunately, the installation bottleneck is getting solved.
However: The problem with players in the installation phase is that they need to put their workforce into work. They need utilisation. And they know their own product, their one solution. Which means: they are not credible in providing the right merit order of upgrades that is specific to an owner’s building.
Should it be solar first? Well, if you ask the solar players, of course !
Should it be heat pump first? According to heat pump players, you bet !
(in reality, in many instances it should be insulation or windows or something else first, so you see the problem)
The operating phase does NOT capture the physical substance
In the operating phase, we have utilities, meter operators, building management systems and price comparison portals who all have a role to play to manage the building’s power consumption. However, it’s limited. For one, their help with heat consumption is not on par. Secondly, the monitoring of power consumption is spotty, at best. Smart meters are still not widely adopted in Germany and larger parts of Europe (I researched and wrote about smart meters almost twenty years ago …). And finally, even if you monitor the level of power consumption, IT DOES NOT TELL YOU ABOUT THE BUILDING SUBSTANCE and why the losses PHYSICALLY occur.
Bringing technology into the assessment phase is the missing link
That’s why there is the assessment phase. It covers the things an owner needs to know about the physical substance and ongoings in her or his building. Unfortunately, the assessment phase is equally (if not more) plagued by skilled workforce shortage. Today, Germany has ca. 7’000 – 8’000 people trained, certified and active to conduct a building energy audit and create the roadmap for upgrade measures, of which many only do this part-time. These highly trained people are called “energy consultants / Energieberater”. They often come from a first life as an architect or engineer, and take on tens of hours of additional training to be certified as energy consultant and be allowed to create the audit documents.
The audit documents are no slouch. They contain 20-40 pages with the building’s specific physical substance, followed by complex normed calculations, followed by an individual upgrade roadmap (what I call the merit-order of upgrades). The merit order tells the building owner what is the most efficient upgrade to do now, to do after, to do later.
And apart from being just really insanely useful – the audit documentation also serves as a purpose to access grants/subsidies and bank financing.
That’s why the building energy audit is really awesome, but complex.
And unfortunately, that’s why a lot of owners want it. If you are a consumer, and you want your single family home audited, it’s quite normal that you wait 6-9 months for your building energy audit.
Now imagine, you are the owner of ten buildings, or hundreds, or thousands. You need a scalable solution – a one-stop shop that uses technology.
Enter: Enter !
Enter has established itself as the leading building energy audit provider in Germany, with more than 3’000 projects completed as I write this. They grew in 600% in the last months. Their secret to achieving such scale recently is that they built a suite of proprietary technologies that allows them to capture building data from owners extremely hands-off, and run the audit calculations and merit-order supported by technology. In the end, it even helps the owners access grants/subsidies by running the process for them. Thx, technology !
What’s equally important to succeed in this market, though, is the positioning and model.
Enter has no workforce to utilize. It is not an installer. As I wrote above, there’s plenty out there that Enter partners with and there’s great synergy. But because Enter doesn’t have to put their own workforce into work, it has no conflict of interest. Enter is in the position to freely recommend any upgrade measure in the order that’s best for the building owner.
Because Enter has a 360° view of the physical substance of the building, its recommendations are based on facts specific to the building. Not beliefs that maybe solar is always best…
And all of this allows Enter to be the preferred partner to owners as well as lenders, since they look to Enter to underwrite the upgrade measures for financing and the actual execution of those measures in the end.
I am convinced that with this round, Enter is scaling not just to a huge financial success – more importantly, Enter is now scaling to decarbonize 20M homes in Germany as fast as possible because owners now finally know what their building’s physical substance is, and what’s the right merit-order of thermal upgrade measures.
a16z published a new post last week about their view on Vertical Operating Systems. In it they argue that:
almost all vertical operating systems start with a best-in-class wedge product that creates a system of record around the foundational customer unit. And then, from that initial foothold, the company launches other products and services that capture more and more of the employee workflow.
Good timing, because I believe we are seeing the Vertical Operating System for SMB contractors in Europe. And it coincides with a major new round announced today.
While a wide range of solutions for the financial stack of SMB contractors has emerged in the US – eg. Adaptive, Briq, Flashtract or Flexbase – the European markets had remained starved for strong venture-backed solutions.
Until Graneet’s new round.
What is Graneet ?
Graneet is the financial operating system for small and medium-sized construction companies. Contractors manage quotes, invoices, project financials and margins, cash flows and payments with Graneet’s software. Key financial workflows are automated, and numbers become transparent across dashboards in real-time.
A useful analogy from other industries is Pennylane – in a super dumbed-down way, you can think of Graneet as the Pennylane for Construction.
Why verticalized ?
If you wonder why a horizontal or multi-vertical financial OS such as Pennylane wouldn’t be better suited to also serve construction, I don’t blame you. Fair question, especially if you have never looked into the specifics of the construction sector.
Construction contractors deal with a bunch of specifics in their financial workflows that no other sector has – not retail, not ecommerce, not logistics, not manufacturing etc.
For starters, financials are arranged around projects, not orders or products or months (I call this the smallest unit of aggregation, SUA). Contractors need to manage their margins on a project level, i.e. their SUA is a project. A project can consist of many milestones and pay-gates, which themselves are made up of orders, applications for payments, partial invoices etc. A financial-OS needs to facilitate these workflow specifics around quotes, (partial) invoices, and margin reporting.
Secondly, projects – and its associated revenues, COGS and cash flows – often run over multiple months, not just one or two. This means that you carry bookings for a long time, while having partial realized revenue in various different months booked against the project, while you carry the remainder of the booking. A horizontal financial-OS which allows an SMB to run their firm around a monthly view only will significantly distort a contractors financials and unit economics. A financial-OS for contractors needs to allow managing your financials not only around projects (see above), but additionally manage across 3+ months of realized revenue, to manage the contractor’s business model.
Thirdly, construction quotes and invoices contain a lot of specific terminology and units. Both quotes and invoices typically have to be broken down to super-specific line items, and line items contain different quantity-units – eg. one being hours, a second being tons, a third being cubic meters, a fourth being distance, and so on. While in retail, ecommerce or services, the units usually are the same and a quote contains anywhere between 1-5 line items, construction contractors routinely offer 50 to 500 line items because it is required by law or by procurement policy from their clients. A financial-OS needs to allow easy quote/invoice creation, editing and summing with multiple different units and hundreds of line items. Also, not unimportant: only a consolidated view of invoices, receivables, project execution milestones (!) and payments will allow instant invoice-financing.
This list of construction-specifics could go on for much longer – just to mention other construction-specific items:
ESG appendices to quotes/invoices and supply chain documentation
applications for payment
insurance certificates, compliance documents, liens
early payments discounts and materials financing solutions
procurement of materials from preferred construction suppliers and inventory-keeping
…
This is the reason why horizontal players prefer going after many many many other verticals first, before they adapt their software to construction. And it’s why a vertical financial-OS wins.
Why not a US venture-backed financial OS for Europe?
US and European SMB markets have fundamentally different mechanics:
Expense cash flows: US market is credit-card (and cash) based, Europe is wire-transfer based.
Labor and payroll: US SMB contractor market has a lot of roaming/freelance and at-will labor with weekly or biweekly payroll, not seldomly on a cash basis and hourly wages. European SMB contractors have a higher share of fixed employees with fixed base salaries paid via wires, and more often than US on a monthly basis.
Compliance documentation: US and European markets can both be heavy on compliance documentation, but what documents are required differs massively. For example, US contractors need to show a lot of insurance documentation, while European SMBs often need to provide documentation from their workers that they were paid minimum wage. Very different set of compliance docs.
Again, this list goes on. Both are mature markets, but organize their financial workflows quite differently. That has implications for the transferability of software from one continent to the other in the early days.
(NB: UK workflows tend to be closer aligned to US markets than mainland Europe, but still different to US)
Cool – so why will Graneet win ?
Amazing product, amazing team, amazing growth. Three out of the biggest criteria for any VC investor.
Graneet’s month-on-month growth has been consistently at 15%+ in 2022.
But such growth can more often than not be bought with expensive and un-scaleable sales motions.
Therefore, let me focus here on a fourth criterion that is quite unique to the firm and prepared Graneet for exceptional scale: its sales efficiency and product stickiness.
When I looked at their 2022 numbers and compared them to best-in-class SMB SaaS benchmarks, this came out:
Sales cycle: Top percentile. To give you an idea: Graneet is measuring its sales cycle in days…
WAU/MAU: this might be the craziest number I have ever seen for an SMB SaaS. Top percentile doesn’t even do it justice. Suffice to say – customers have an insane recurring usage of the software.
Sales burn multiple: Top decile, profitable
In Graneet’s case, what blew my mind is the quite insane customer love coupled with a motion that scales and a very high dollar retention in the SMB segment.
It’s rare to see a vertical SaaS with such excellent metrics poised for capital-efficient growth.
Now that Graneet has loaded up on cash, with a product that is loved by a wide range of contractors, a European market starved for a great venture-backed solution, and possibly one of the greatest vertical SaaS investors of all time (Point Nine), I believe the firm has the ingredients to scale very rapidly.
Over the next weeks, I will be sharing a series of posts explaining our view why the 3D design stack is not just ripe for disruption – but why it is already in the middle of it – and what our thesis is.
To begin, it’s important to pin down the problem.
And no better way to show you the problem than visually.
Incumbents’ 3D and CAD software have a huge problem. They are built on older stacks. When I look at it, I see clunky, swiss-army-knife style software that requires a ton of training time and locks users in by making any switch away as hard as possible.
On top, legacy CAD firms treat users not like kings. Prices seem to get hiked, license structures changed, service seems mediocre.
And users take notice. It seems they took notice for a long time. And they vent. A lot.
Legacy CAD firms don’t seem to care much. It looks like they double down on their strategy. By buying, and buying, and buying more independent software firms, sunsetting software, and closing it all off.
The disconnect seems real. Employees at legacy CAD firms seem to love their life and strategy.
Users not so much.
Meanwhile, the UI/UX design stack has delivered the blueprint for how the exact same situation has played out there.
One of our super exciting young bets of the last year had been Lun, who today announced a new €10M round with Norrsken, Lowercarbon and Partech, which they raised ultra quickly on the back of assembling one of the best management teams in the space in Europe and executing a super unique product (more on that below). We are proud to have led Lun’s pre-seed just 12 months ago.
For a bunch of folks out there the above Venn diagram might not make sense at first. What do design and workflows have to do with heat pumps. It‘s all about buying small shops, training some people, isn’t it?
No, it isn’t.
A lot of companies approach the thermal renovation opportunity across Europe from an Enpal-like angle (basically super-heavy asset-financing model). There’s Thermondo, 1KOMMA5 and many others with slightly different but fundamentally similar takes across Europe etc.
If you really break it down to first principles, these models resemble asset-financing businesses where they take some workforce constraints out. The workforce aspect is secondary though – when you read the P&L and Cash Flow Statements, you see that these models are asset-financing with some bolt-ons left and right.
The not so attractive part with that approach imo is (1) this is a model born out of a low interest environment, and the cash economics of such models are just not what I look for in firms (2) you cannot be a credible arbitrator of a merit order of renovation measures, cause you know, it has to be always eg. solar first (even if the building would benefit from another thermal upgrade first) because that’s where I need to put my OWN workforce into constant work and (3) if you look at reducing our CO2 footprint on the grandest of scales – we are talking 100M+ homes in Europe as fast as possible – these asset-heavy models just cannot move the needle. Only a software play that enables hundreds of thousands of installers can reach that scale and reach it quickly.
Therefore, my thesis is different:
100M+ European homes need thermal upgrades as fast as possible.
Not every building has the same merit order of upgrades. Some have the biggest initial impact from insulation first, some from heat pumps first, some from solar first, some from windows first, and so on.
One of the fastest growing upgrades will be heat pumps (the other will be insulation, by the way, but more on that in another post soon).
Heat pump upgrades suffer from a massive, massive demand-overhang / under-supply.
Material supply will get fixed more quickly than labor supply. Labor supply will remain the largest constraint.
Spec’ing heat pumps is more complex than rooftop solar or fossil boilers. Lots and lots of variables go into spec’ing it right, and the consequences of being wrong are much more economically and thermally significant than for other upgrade measures.
Installing heat pumps is also much more complex, and information is harder to come by than for rooftop solar.
Fortunately, we have a large pool of skilled craftspeople who know how to spec and install heat pumps.
But: they are severely capacity-constrained. 50% of their time is spent on admin tasks that software can help fix: on-site visits, writing quotes, writing invoices, chasing customers to convert and chasing them to pay, and procurement – and collating the various data points required to spec the heat pump, some from open sources and some from the building owner directly.
And: only 25-35% of quotes written by heat pump installers convert into orders. Software can help with this, too.
Bonus: If you build the industry-standard software, a new generation of installers that gets groomed in the coming years can be powered by this software without being locked in to any single employer, and be their own employers. The term “business-in-a-box” applies.
Therefore, my thesis is that a B2B installer-centric design and workflow softwarefor heat pumps which helps with spec’ing and all the tedious overhead work and customer-centric and supplier-centric workflows in itself can free up 50% of installer capacity (i.e. double the current capacity) across Europe. And do this fast and at scale.
In other words: Scaling heat pumps to mega-scale is a design and workflow problem (because it requires skill and is highly varied).
There is precedent and a playbook to get inspired by from the US in rooftop solar – Aurora. And for heat pumps, this formula is even more relevant due to their complexity.
And that’s what Lun does for European heat pumps. No marketplace, no asset-heavy workforce – just amazing design + workflow software to enable installers.
Lun has assembled one of the very best climate-tech and software management teams in Europe. The next year could be fun, now that we stocked up with cash 🙂
P.S. some resources for my readers who love physics and who’d like to become heat pump geeks:
77% of freight transported in the EU is carried by trucks. That leads to about 6.2 million trucks circulating through the EU.
All of this truck freight directly links to gross domestic product. Less drivers, fewer road freight, less goods from the upstream supply chains arriving at their downstream destinations.
And no, increasing rail- and ocean freight do not solve this issue for us. The vast, vast majority of road freight is carried for less than 150 kilometers, which already accounts for the rail and ocean freight which carried goods further upstream.
The truck driver shortage is a big problem
In 2021 we were ±460’000 truck drivers short in Europe + Turkey. These unfilled jobs accounted for ca. 10% of the entire truck driver jobs in these markets.
The breakdown (ca.):
UK: 80’000
Poland: 80’000
Turkey: 80’000
Germany: 70’000
Romania: 70’000
All others: 100’000
And the trend was not our friend: Truck drivers dropped by 10% in UK from 2020 to 2021. 40% of open driver jobs are unfilled.
It’s not like fleet managers do not see this problem (they do !). For example, in UK in 2021, drivers wages increased by +18% (this is before the Fed began acknowledging the spiralling inflation), but they do apparently not increase enough to rebuild the dwindling truck driver supply.
… getting much worse very soon
In 2021, 230’000 drivers had drivers license attestations from non-EU countries, of which many are Ukrainian or Belarusian. According to IRU, on top of the shortage we saw at the end of 2021, an additional 170’000 truck drivers may have left the jobs due to the Ukraine war since beginning of 2022.
If you add it up, just based on this, we are probably short 640’000 truck drivers in Europe + Turkey today.
Now here’s where it gets spicy.
The average age of truck drivers in Europe and Turkey is close to 50 years. 25% of truck drivers are older than 55 years – almost 1 million truck drivers !.
If you add this up to the already existing gap, we are looking at a 1.5M+ truck driver gap in the coming years !
The #1 issue cited by folks who know the road freight industry are: working conditions !
Fixing working conditions for truck drivers can be broken down into four pillars:
Increase time at home (very interested in such solutions !)
Decrease time pressure
Improve loading/unloading conditions
Improve hygienic conditions
If you fix all four, you might be building a very big business.
Specifically, I am looking for founders who use technology to source and train drivers, provide best employment, maximize earnings for drivers and, in doing so, increase time at home (eg. remote solutions instead of autonomy solutions).
(What’s missing on this list is level 4/5 on-highway truck autonomy. Every automotive engineering expert I am speaking to tells me the same thing: not in the next 10+ years. Truck drivers will be here to stay is my bet)
The European Central Bank raised interest rates a few weeks ago again, bringing the main lending rate to 3% and further closing the gap to the Federal Reserve.
I hadn’t had much time to write about it earlier, but I am deeply concerned from a societal perspective. How will thousands of families in Europe (and USA) continue to afford owning homes? Or is there a relief on the horizon?
(I’ll use German numbers below – many mature markets will show directionally similar dynamics)
50% of households have a net income of €2’900 or less. 90% have €6’500 or less net available.
I was interested to analyze how much home households across the spectrum were able to afford in a low-interest rate environment (1% mortgage interest rates were commonly available until 2021 in Germany).
Then I wanted to understand how that affordability of home prices changes as interest rates increase.
For this analysis I used rules of thumb and mortgage assumptions commonly used in Europe:
2% initial annual repayment of principal (norm in Germany)
at least 60% of net income needed for non-mortgage living costbut max. €8’000 per month (banking rule of thumb in Germany)
using a Financing of 100% of home price; transaction cost paid from cash savings (used for simpler calculation; more cash = more home)
Here’s how it looks:
As you can see, the affordability drastically decreases with rising interest rates. Between a 1% mortgage and a 6% mortgage, the home price that a household can pay is reduced by 63%.
If you now overlay the distribution of household incomes, you can make a very reasonable assumption that 90% of homes were purchased for €1M or less, the other 10% were possibly purchased above that.
Here is where it gets (depressingly) interesting:
Flip the perspective, and ask yourself – how many households were able to afford a certain home price in a 1% environment (aka market size for that home category) and home many households in a 6% (adjusted market size for that home category):
For a home that is put on the market for €460k:
50% of households could afford it in a 1% interest environment
3% of households in a 6% environment (you read that right !)
That means you have a 47%-points smaller market size for a home marketed at €460k between 1% and 6% interest rate.
And this is where it becomes a societal issue:
What you see here is that the change in market size happens in the low-mid home categories, by as much as -50%. While in the homes that households with upper-end net incomes can and will afford, you have practically no change in market size (0.5%).
Now, the first conclusion one could draw here is that home in low-mid price brackets could crash in price, while upper-end homes could be more stable in price. That is a reasonable conclusion.
But the story doesn’t end here. Consider the cost of new construction, which is the equivalent of creating new supply in the market:
Unfortunately, construction cost keep escalating. This comes down to skilled labor shortage and availability (and prices) of materials. Regulation doesn’t make construction easier, but more complex. Thus, experts expect construction cost to keep rising.
Which means, we cannot expect new supply to flood the market and create relief in the form of widely available, cheaper homes – at least not immediately.
And this is the very concerning part:
A large part of households signed mortgages at ±1%
When they need to refinance, will their homes be in a price bracket that could come under price pressure, as the market size decreased by as much as 50%?
Even if that doesn’t happen, can they refinance the same debt (from the home price paid in 1% environment) with higher interests?
The silver lining that happened in 2008 in large parts of the US was that home prices came down. The difference is that there was a very steady stream of new home construction in the US available back then which put homes under pressure. I don’t see this happening in Europe this time around.
Therefore, my fear is that we are already in the process of not just keeping new generations from owning homes (that’s already happening), but that former owners will be forced to become renters as institutional buyers are the ones swooping up the low-mid price brackets of homes with available liquidity.
I dug deeper on how recycling and circularity of construction & demolition waste (CD&W) might or might not help fix the raw material shortage we are running into.
To understand this, I had to understand what waste is generated on construction sites, what the CD&W is made of, and where it goes.
(I focused this post on the US market; European and Australian markets will have different specifics but will be directionally same ballpark)
Construction & demolition waste increased by +340% compared to 1990. While – interestingly – municipal solid waste “only” icreased by +40%. We create disproportionately more construction trash than household trash.
The big 6 components that make up construction waste are:
Concrete
Asphalt pavement
Wood
Gypsum (from drywalls)
Bricks and clay
Metals (steel, copper, aluminum and others)
All other materials fall into the “any others” bucket. An example of “others” would be asphalt roof shingles.
Concrete makes up 70% of all construction waste. 22% are asphalt pavements and wood. Bricks, clay, gypsum, metals, and any others together make up 8%.
When it comes to “where does it all go”, brief time for definition:
Re-cycled means something is re-used in (broadly) the same way as before. For example, asphalt pavement re-used to build new asphalt pavement.
Down-cycled means something is not just dumped but used in an energetically lower yield than its prior use. For example, burning wooden doors as fuel.
Un-cycled means something is just dumped in a landfill. Here it’s not even used as fuel or as an aggregate to fill a hole for a new site.
To the numbers:
20% are re-cycled. The majority here is asphalt pavement. Some concrete is recycled. Some minor recycling of wood and metals.
56% are down-cycled. Vast majority is concrete (mostly re-used as aggregate). Some down-cycling of pavement, wood, gypsum, bricks and clay.
24% are un-cycled and just dumped. Again, a good share of this is concrete, but surprisingly much is wood. The majority of bricks, clay, gypsum and any others goes here. Reasons are diverse, but two of the bigger reasons are that separation is too costly or that adverse materials are contained.
What all these numbers tell me is that creating a stream of homogeneous supply of recycled construction materials is really difficult – and really key. The two biggest levers to improve our recycling rate are (1) concrete and (2) wood.
For concrete, someone has to fix the separation to be economically scalable.
For wood framing, I have seen approaches to remove nails – that is generally useful, but holes remain and since wood ages, the structural integrity decreases. That’s why it gets down-cycled.
For wooden door frames, hardwood floors, and similars: they have visual properties that lead to in-homogeneous supply at scale (every door looks a little different)
Who’s fixing construction waste at scale using technology? I.p. looking for robotics and software-led approaches
Had a chat with a founding team last week who provided a sharp observation:
“Construction is full of these hidden billion-dollar problems”.
I’ve made the same observation over the past years of investing in early-stage ventures with Foundamental.
But these opportunities are only hidden to construction-outsiders, not to the insiders. That makes them tougher to access for stellar founders without years of working on a site.
One common fix is when second-time founders – after they had made significant splash with a first exit – take 1-2 years and spend time on construction projects and in the supply chains themselves. This happens more often than you might think, for example:
Hence, want to focus this week’s post on some of the most unexpected and large hidden truths in construction I’ve learnt about as an outsider myself over the past years to give y’all a headstart of where to dig deeper. Not meant to be an exhaustive list, nor to provide the in-depth. Just to mark an X on your map, not going to do the digging for you. Below are some of my more favorite hidden opportunities I learnt about over time:
The mess on construction sites can be really that bad
1/ Document controllers
A lot of project sites still rely on offline paper to record information, allocate work, and keep an audit trail. Documents get wet, damaged, lost, confused. From customer invoices and signoffs to fuel expenses, site risk forms, and everything in between.
Did you know that there is a job (in larger projects) called a Document Controller? This person’s task is to make sure that the thousands of documents flying around the site are properly collected and distributed among the other members of the team, eg. for signatures, approval and validations, and get returned on time and with the proper input. This role exists because of the crazy amount of paper you need to run an auditable construction site.
2/ Thousands of versions of a project schedule
A construction schedule is the planned sequencing and timeline of tasks and milestones in a project. It can easily have tens of thousands of line items in larger projects. It’s not a secret anymore that optimizing schedules is a billion dollar opportunity, with several startups active in this opportunity space but with different angles and approaches. (eg. Alice, Nodes & Links, Foresight, nPlan).
What is less of a known fact to outsiders is that there are thousands and thousands of version of a schedule, because changes happen all the time, which requires all subsequent dependencies to change as well. The amount of files and documentation for collaboration around the schedule therefore is a nightmare.
3/ Estimators for materials take-offs
“Take-off” is an Americanized expression for assessing material volumes (the bill of materials, BOM) and the associated cost of those volumes for a construction or renovation project.
In the US market specifically, it’s quite common to have an (often external) “estimator” look at the blueprints or models of a project and conduct the estimation of volumes, and provide researched or experienced-based cost estimates.
This process is ultra tedious, easily taking a lot of man-days from a professional estimator.
Construction = supply chains + working capital
4/ Sub-contractors deliver materials early
Let’s continue with one of my craziest favorites. Sub-contractors around the world very often offer the materials associated with their labor in their quote to the general contractors. This allows them to choose materials most comfortable with (within specs), and most importantly, skim some materials margin as additional income.
I’ve learnt that in some of the European low/mid/high-rise markets, sub-contractors have a habit of delivering their quoted materials weeks earlier than their labor starts.
Why? Because it allows them to invoice the material to the general contractor and get paid for it earlier than their cash clock for their labor starts. I would not be surprised if they also try to receive more favorable payment terms for their distributors or manufacturers, thus optimizing their cash cycle. I can’t blame them, since sub-contractors get squeezed constantly for cash from both sides.
5/ Applications for payment
In many markets around the globe – especially Anglo-American markets – applications for payment (AFP) documents are a standard with which a (sub-)contractor provides an overview of (partial) labor and material that has been rendered. With this document, the contractor requests to be paid.
The customer (eg. a general contractor) passes the AFP around internally to the colleagues best suited to provide judgment whether the work has been rendered satisfactorily.
There are hundreds of millions and millions of AFPs flying around the construction markets every year. Its significance is so high because it not only expedites payment – it enables an important quality assurance workflow through which the client is forced to judge acceptance of work and material.
While there are several standard forms for an AFP, reality is: millions of documents fly around.
6/ Payroll companies and crew leaders
This is a UK special, but might be happening in other markets, too.
In the UK, it’s quite common for construction companies to hire flexible construction labor, often entire crews who keep working and roaming together.
The UK has recognized this many years ago and implemented CIS – the Construction Industry Scheme that regulates how payments, payroll and taxes in construction ought to be made.
To run such payroll creates a headache for construction firms who employ roaming crews. One common practice has therefore become to have specialized construction payroll firms. For example this one (among many others in the UK).
Now here’s where it gets really interesting: Among such roaming crews, often times a crew leader (sometimes called a “baron”) can emerge who markets “their” crew to contractors and their projects. In the payroll processing, they take a cut of the entire payroll (regularly 10-20% of the entire net payroll) and have the remainder paid out to “their” crew. Interesting, huh…?
7/ We speak not enough about aggregates
You might have heard that the world’s consumption of raw materials is set to double by 2060.
But when you hear “raw materials”, you might think oil & gas, metals, coal, wood, or even agricultural products.
Very few people think of sand, gravel and crushed rock – the so called “aggregates”.
55% of the entire world’s raw material consumption in 2060 will be sand, gravel and crushed rock. A material heavily used in concrete, bedrocks and other construction applications.
The construction supply chains would not work without third-party trucks and drivers. Since I mentioned about aggregates above, let’s use that as an example:
Customer orders aggregates from a quarry
Quarries and 1 and 2 says “sorry all sold out right now”
Quarry 3 says “yep send a truck, he can load up”
Customer orders an external truck – often via a local broker
The broker and truck driver know all the quarries around the town
But to get onto the quarry, you need to get through a gate
The gate has physical access keys in cheque-card-format
And guess what: That’s what the truck drivers have under their sunscreen – 10-20 access keys to all the quarries in town they drive regularly to
And who keeps track of these keys? Hell if I know…
Construction is ultra-relevant – even before considering CO2
9/ The #1 reason for fatalities are falls
Falls are the leading cause of deaths among construction workers. Not electrocutions, not being struck by or caught between objects. Falls.
Consider this crazy stat: According to Safety & Health Magazine “over a 45-year career, a construction worker has a 75% likelihood of experiencing a disabling injury and a one in 200 chance of being fatally injured in the job.”
10/ 220 million people work in construction worldwide
Leaving this here feels a good way to end this list. Construction is insanely relevant. Not just in terms of its share of GDP (between 10-13% worldwide) or its total market size (somewhere between $10-16 trillion). But for the employed lives it impacts. You want to have impact as founder, don’t look beyond the construction sector.
Artificial intelligence (AI) seems to be the latest catch-all that venture capital investors are in love with – following “FBA aggregators” in 2020, “Web3” in 2021 or some fundamentally stupid “climate” stuff that also got backed in 2022.
But I think applied AI can equally transform vertical B2B sectors.
One such huge opportunity I am seeing is AI-enabled vertical search in construction and architecture, specifically vertical search for light building materials such as floors, tiles, textiles, paints, decor and so on.
When I posted this on Linkedin the other day, a discussion erupted about concrete, cement, steel. Those are heavy “in-the-wall” building materials, which in the mature markets have very established supply chains and market dynamics (that’s different in emerging markets).
For vertical search, my observation is that light “outside-the-wall” building materials make a perfect fit. Categories such as paints, floor, stones, fabrics, glass, bricks, decorative metals, plasters, textiles have more than 250’000 potential EU-based suppliers, millions of customers and millions of SKUs.
Most importantly, the way the discovery, selection and purchasing happens is diametrically opposite to heavy building materials. Consider my framework here:
Framework: Purchasing decisions in light building materials vs. heavy building materials
Buyers of such light building materials know what they want as they have a vision and concept, but need to find specific SKUs.
And for these light material categories, purchasing decisions are made not just on performance and price, but additionally on haptics, optics and increasingly ESG considerations.
Therefore, the industry hotfix practice for the last 50+ years has been to (a) meet at trade shows and display samples and (b) send around samples in large quantities during the early discovery processes. Both practices are not scalable for millions of SKUs in light building materials.
That’s where AI-powered vertical search comes into play.
My dumbed down thesis is that for every “Amazon” in light building materials (eg. Material Bank), which is logistics-enabled and strips manufacturers of their market power, there must be a software-powered “Shopify” that enables manufacturers (eg. Mattoboard, Stylib and others)
The one reinforces the other.
Now applied AI becomes increasingly ready to allow manufacturers to organize their vast SKU catalogues at scale, while it allows buyers to discover one very specific SKU they need for their project without going to a trade show nor getting more samples. At scale !
Hence why I am keen to back founders building the AI-powered vertical search for (light) building materials in mature markets.
I have a bit less time this week for a longer researched piece. Since we just kicked 2023 off, thought it would be interesting to share a few spaces and models which got me excited in 2022 across the AEC-Tech world, and where they might be going in 2023.
(example investments I mention below include both first cheques and follow-ons)
The $20B acquisition of Figma by Adobe in September brought multiplayer design back on the menu of many investors. We believe that multiplayer design and CAD – both in 2D and 3D – will be vertical. Demonstrated already by the eg. ntopology, we also believe that retooling the architecture, engineering and design stack with multiplayer solutions, and building a n:n API infrastructure around it, is a Decacorn frontier.
I am EXTREMELY keen to do do more in retooling the AEC design space with great software and highly differentiated technology in 2023.
Enabling thermal renovation
Building owners operate ca. 69 billion square meters of floor area in North America and Europe. And all that space needs heating. Gas supply came under shock in 2022, and the result was an extreme price volatility in fossil fuels.
Well, in Europe, 80% of buildings are heated with fossils – gas and oil (and in some parts of Europe even coal).
Additionally, the EU is introducing ever-more ambitious decarbonization targets and regulation for building owners.
Hence, we made multiple new partnerships early in Q1 2022 with amazing teams who fix the rapid thermal renovation and upgrading of the European building stock.
Example investments from 2022 are Baupal, Lun and a company I cannot announce yet but which is building the one-stop shop for retrofitting insulation across European buildings.
This space has developed rapidly in 2022 with several amazing teams. In 2023, I am keeping my eyes peeled for various not yet-addressed opportunities in this space, one of which is fixing the financing journey for B2B and B2C customers in thermal renovation.
Replenishing the skilled labor pool
The West is running into very deep issues with regard to skilled labor supply. I plan to run a detailed research piece about this in the coming weeks.
We see 3 levers to fix this issue:
Making existing labor supply more fungible and efficient. Fixing the candidate journey to allow super efficient recruiting admin processes. And providing software tools to do repeatable jobs repeatedly.
Upskilling domestic labor supply. Providing efficient (re-) training offerings in person but strongly tech-enabled. Can come with process standardization for employers, eg. through academies-as-a-service.
Qualifying, upskilling and providing mobility for foreign labor supply. There hundreds of thousands of exceptionally skilled – and socially aspirational ! – people outside of the Western markets. Solutions in this space make those mobile.
Very keen to dig deeper into blue-collar academies, tech-enabled full-stack employers-as-a-service, and payroll and earned wage access solutions for construction and renovation, among others.
Automating workflows with advanced robotics
Same fundamental market gap as on the skilled labor shortage front, but solved differently.
Very excited about advanced robotics for construction (renovation also, but much more difficult due to lowest standardization and controllability of surrounding environment).
When partnering with robotics founders, I am looking for
Highly differentiated tech with strong software components
High degree of automation on the whole workflow (not just a small part of it)
Looking for spatial awareness. More defined by machine vision than traditional robotic automation
Extremely savvy go-to market motion aligned with how construction likes to work
We are extremely bullish on robotics for construction, and will look at every deal that fits our above criteria.
Providing one-stop procurement options
Construction is moving into the teenager stage of digitalization, and part of that is because enterprise customers are adopting project management softwares and one-stop shops for building materials. From both ends, we are seeing how more and more integration opportunities emerge, which are resulting in even bigger one-stop shops. Infra.Market from our Fund-1 portfolio is a great example of this, and we see this trend go around the world.
Example investments from 2022 are Brick & Bolt, GoCement, Latii, an unannounced one-stop shop for cross-border building materials in Germany, and an unannounced one-stop shop for sub-saharan Africa, and an unannounced Zetwerk for Mexico.
Super keen to do more in this space, and in Europe specifically, procurement softwares that connect domestic suppliers with domestic demand on top of which they load software and procure-to-cash workflow automation.
Un-tightening the cash cycle
300 basis points on central bank rates make a huge difference at wall street, and in the real economy. We have seen a rise of receivables and payables at contractors in 2022, across enterprise and SMB.
All the while, in construction, suppliers have always been the secret working capital banks of the industry. Now with increasing cash cycles, this reaches real limits.
As margins come under pressure as well, the ability to control my contractor business and my projects regains a premium.
With these tailwinds, we have been excited to back financial operating systems, materials-price hedging solutions, and automated access to working capital for contractors and suppliers.
Recent example investments are Tazapay, Earlytrade, Graneet, Adaptive and an unannounced commodity price-hedging platform for general contractors.
Very keen to partner with more and more differentiated avatars across the world un-tightening the cash cycles across AEC.
Building vertical-search infrastructure
And a last big thing going into 2023 on my personal radar is vertical search within AEC. The release of OpenAI’s trained ChatGPT model in November caused a major re-buzz around what AI will be able to do.
Within my scope – architecture, engineering and construction – I see major applications of AI for vertical search. By vertical search, I mean for example that we have millions of stock-keeping units (SKUs) in our industry, over tiles, paints, roof shingles, concrete, steel variants, fabrics and so on. And these SKUs are described by dozens of properties. It is a major pain for manufacturers to manage their product portfolio, let alone make it truly searchable for their customers – designers, architects and contractors.
At the same time, manufacturers keep pushing the boundary on moving closer and closer to their end customers, but without the infrastructure to allow direct B2B shopping.
I am super keen to partner with founders who build vertical search infrastructure, ideally with design-stack components and webshop components. Most excited by distribution as infrastructure and API plays.
An example investment we did in 2022 is Mattoboard.
Hit me up if you are building in any of these spaces, or other frontier spaces in AEC-Tech. I love to be ahead of the curve.
In the past week we announced our new round in Snaptrude, which we were fortunate to co-lead alongside our friends at Accel.
Snaptrude, which is like the Figma for AEC, is one example of highly talented product teams re-inventing the 3D/2D design stack.
Like Altaf, we believe in a multi-player and open (!!!) future for the future design stack. And that’s a transformative leap, because designers today are stuck in hundreds of walled gardens from legacy players.
Architects and engineers using CAD spend more than 700 hours per design with their software stack. And they have to live with insane versioning.
Here is a super crude example of how nuts the design versioning in AEC is:
1 project, 1 model (this is a simplified example; the reality is much worse)
10+ collaborators
100+ softwares in use across the collaborators
1’000+ versions
10’000+ files sent around
100’000+ file uploads
… for one project
And the industry is made up of 607’000 engineering/archi firms. The market leader worldwide has 0.3% market share (you read that right). I wrote here about the market structure.
In other words: multiple firms collaborate on the same project.
And interestingly, this fragmentation is increasing. In 2015, 52% of ENR firms had 50+ employees. Four years later, it was only 49%. The super small firms also lost some share, down from 15% to 13%. It is the “mid-sized” engineering/archi firms with 10-49 employees who are gaining share, up from 33% to 38, and rising.
Meaning, mid-size firms – which are still very small by any other industry standard – are becoming the norm.
On top of this, we can observe a structural shift over the past years towards multi-disciplinary firms, i.e. engineering/archi firms who house the different disciplines required to engineer a project internally. Which does not eliminate the need for collaboration, but inversely should increase the speed of adoption for collaboration solutions within the same firm.
The rise of mid-sized and multi-disciplinary engineering/archi firms is relevant for the design stack as mid-size firms should be easiest to penetrate and gain adoption with. Small firms tend to not have the overhead staff to deal with adoption. Very large firms come with the corporate BS and bureaucracy which leads to slow adoption and long sales cycles. I am most excited about the rise of mid-size firms.
Lo and behold, this adoption thesis seems corroborated by this chart – how fast the different firm sizes adopted BIM over the past two decades. The middle line is the mid-sized firms. I wish I had sales cycle data to plot against it, but my money is that sales cycle with the large firms will be 3x longer than with mid-sized firms.
So, we can see that (1) multiplayer is not just a fancy trend, but a structural need with structurally-supported tailwinds and (2) selling multi-player design tools to SMBs in AEC gives you a growing market within the fastest-adopting segment.
Now, luckily, this opportunity is further amplified by a third factor, namely (3) CAD users really really really do not like the legacy solutions.
And well, the incumbents tell us quite frankly on their own website how they see the future of the ecosystem. What they call “moving interoperability forward” I would call “creating a walled garden with some peek holes”.
Meaning: The incumbents telegraph their vision of more walled gardens, not less – whereas the market is structurally moving into the multiplayer direction.
We have seen this same structural shift play out with the 2D graphic design community. Adobe legacy solutions, Figma and Miro transforming it. And eventually, it ended here:
So for me, it is crystal clear in what place the 3D/2D CAD design stack will end. It seems to me quite clear the legacy authoring tools – which are solo-player, 1:1 – will try to expand into some multiplayer functionalities and some value-added functionalities that we see develop in other parts of the market, such as eg. ESG analysis functions from Cove.tools. But they will ultimately fail in innovating fast enough, due to the closed nature of their approach in a market that will now favor open tools structurally more quickly.
As a result, pure-bred multiplayer authoring tools, that incorporate versioning, commenting, collaborative editing (minus true branching as you would have on a git), are continuing to gain adoption in specific parts of the design market. Namely Snaptrude, Rayon, Qonic or Arcol.
And these multiplayer tools – equivalent to a Figma – will incorporate more open functionalities and integrations in amazing user interfaces. Now, as transformative as they are, one restriction will remain: they are 1:n tools, meaning all players will ultimately have to author in the same tool. This is the same restriction as with Figma or also Notion.
So, beyond the multiplayer tools, we believe an open infrastructure streaming opportunity is emerging, that will enable all users/players to author in their respective tools, but on the same model. So this is a cross-platform streaming opportunity, that we are extremely excited about. The pioneer and market leader in this 3D authoring streaming opportunity is Speckle, and something similar might emerge for flat documents in AEC (I spoke about the flat document streaming opportunity here)
So, my vision of the future AEC design stack is open. Multiplayer and streaming is what I am excited about.
Last week I traded emails with a McKinsey Partner who counts global AEC software firms among his clients. He was working on a new report to be released soon, and I shared a few editorial thoughts prior to release.
As we exchanged thoughts, one of the topics we had controversial thoughts over was the share of SaaS unicorns in construction vs. other sectors.
There are are 1’100 Unicornsin B2B. Of which 650 (59%) are SaaS.
We have 4’350 Soonicornsin B2B. Of which 2’100 (48%) are SaaS.
And we have 12’330 “Minicorns” in B2B (a loose definition by Tracxn/Pitchbook). Of which 6’980 (57%) are SaaS.
By comparison:
25%ofConstructionTech Unicorns are SaaS (Procore and Built). And it could actually be argued that Built has FinTech promise priced in, and the SaaS component is less driving the equity price.
41% of ConTech Soonicorns are SaaS.
While interestingly, 56% of ConTech Minicorns are SaaS.
75% of ConstructionTech Unicorns are NOT SaaS, while 59% of all other B2B Unicorns ARE SaaS.
Yet founders raise venture capital for as many SaaS firms in ConTech (56%) as in all B2B sectors (57%).
And before you conclude that ConTech SaaS is just a new category and it’s just a time delay. Nope, that’s not it. ConTech SaaS exists for decades, and has been VC backed for the last 10+ years.
So what is happening along the way you think? I have my thesis from the past 4 years investing in AEC-Tech and experience from our early investment in one of the non-SaaS ConTech Unicorns (Infra.Market), but curious to hear your thoughts.
I’ve been seeing a few API-enabled startups in architecture and construction this year. Few of them API-first.
We first partnered with Speckle in 2020, leading their pre-seed raise. Dimitrie and Matteo have created an amazing API-first company distributing 3D data on the most granular object level.
But apart from 3D data, there is massive amounts of flat data in AEC that needs harmonization and easy distribution, such as invoice data, purchase orders, tickets and lots of project level information.
So where do we expect the API-first players in flat data to emerge? What do I look for?
There is 550 billion invoices issued globally per year. 10% of global GDP is construction. So, fair assumption that at least 55 billion invoices are issued in construction every year. That makes at least 165 billion flat documents (invoices, tickets, purchase orders etc.) in AEC. This number is most likely under-estimating the flat documents in construction, given lots of milestones and sub-contracts.
And yet, there is only a comparatively small handful of startups dealing with API-first in flat AEC data as time of writing. Selected examples are Agave, Comstruct,Gryps, or Toric.
So we used our learnings from 2020-2022, and combined them with our analysis of more than 100 well-known API (first) companies from other sectors.
(Here is our API startups market map and database – we made them open-source, so feel free to re-use)
And bam, here are our 10 commandments for API-first in AEC:
#1 Repetitive data objects, used one billion times a year Unified APIs need to process similar data points in same objects a billion times a year.
#2 Done repetitively and manually across systems The data points are used repeatedly in manual workflows with iterative steps, switching systems + channels.
#3 Ability to focus on 1 scaleable property (ideally 2) Either:
one standardizeable data object (eg. ecommerce payment, account statement, 3D model)
one standardized “host” software with market share >10% (eg. Procore, Revit, AutoCAD; NB: SAP doesn’t count due to customized instances)
one super-repeatable customer type who already has tens of softwares and is building integrations with own dev team (eg. Mulesoft, Makini)
#4 Data points can be standardized BEFORE you reach scale You need 1’000 samples to identify repeatable standardization and harmonization. If you need 1’000’000 samples you picked the wrong object.
#5 Already digitally available Data objects exist consistently in software. Implies: pre-requisite is high (legacy) software adoption by suppliers.
#6 Fragmented data providers offer non-harmonized data points Integrations are expensive when point-to-point, fragmented, undocumented, and data is not harmonized.
#7 50%+ of market has no public APIs There is not a dominating software provider that already has open APIs (eg. Twitter, Meta).
#8 hundreds of parties develop applications using the data points Either:
Inhouse engineering teams (eg. compliance, verification, security, eshop multi-payments before Stripe)
Third-party software vendors or startups (eg. FinTechs with Plaid, HR/payroll startups)
#9 High existing software adoption from SMBs/startups (accelerant, not prerequisite) Accelerant because: Low resources to build all integrations inhouse. Hence, high value per customer per software integration offered while initial sales cycles are low and initial frequency is high.
#10 Software procurement done by software engineers (accelerant, not prerequisite) Accelerant because: Very difficult to achieve high sales velocity when procurement needs to pass on the discussion to engineers.
If you are building an API-first play in architecture or construction (or even logistics !) that satisfies as many of our 10 commandments as possible, hit me up !
Over this time, you taught me so many things about you and about business. When you chose to become founders, and chose us as one of your partners, so that together we could build iconic tech firms in AEC – in 3D, design, renovation, energy efficiency and generative AI – it’s been a great ride already.
You were kind to share with me what your previous life’s in architecture taught you.
With which you allowed me a window into your world, at the intersection of architecture meets entrepreneurs meets coder.
Which made me fall in love with you.
Architects, here’s 10 things I love about you – and why you make great founders:
1/ Your product mind: 740k+ of you and folks in your archi-firms are coders. You worked on 3D your whole life. Your whole profession is about figuring out what the users of the buildings you design will need, and formulating the right questions to find out. If you get requirements wrong, your designs fail epically. You are natural product managers, and I freaking love it.
2/ Your founding DNA: Architecture is among the longest tails in business. 1.6M architects working in 607’000 firms across the globe. The largest player (Gensler) has 0.3% market share. Founding is in your blood.
3/ The chip on your shoulder: Architecture is a grind. On average, your employers earn $184k+ per year selling your hours. And they keep $105k of that for themselves. You realize that your hours don’t scale – so you learnt the appetite to scale something bigger than yourself. Your own venture.
4/ Your work ethic and results mindset: You are easily among the hardest working founders I meet. Architecture practice has taught you well. Hard deadlines, short on staff, never any budget for anything. And yet you deliver. Like a great founder.
5/ Your abstraction ability: You get the big picture. Your work always starts with a vision. It can never be bigh enough for you. They teach you to dream big early on. Which makes me love you.
6/ Your attention to detail: And yet, somehow you always manage to sink your teeth into the nittiest and grittiest details – cause without those details, your design will not be built. Founder life.
7/ Your pragmatism: One of the hardest things to get right in venture is to take pragmatic decisions. Architecture taught you this skill from day 1. Your boss asked you to time-box and budget your tasks. Constraints are everywhere. And yet you needed to deliver. Finding pragmatic solutions is the only way to survive as an architect. And as a founder.
8/ Your natural storytelling: All the constraints you face force you to find ways to keep the people around you motivated. I love that you know how to present coherent narratives that make people follow you in spite of constraints. It serves you well in hiring and fundraising.
9/ Your messiah complex: Nobody puts it better than Dimitrie: “We are partly delusionally convinced that the world revolves around the stuff we build and the plans we design. A useful defect to have in order to even consider doing a tech startup.” Couldn’t agree more. Love it.
10/ Your fast learning: You know you are not perfect. Who is? When you found, some of the more common learnings among architects are to delegate better, communicate more precisely, and operationalize everything. The best of you pick it up quickly, and become superhero founders.
I learnt all of this from some of the best of you, who practiced in archi-firms and went on to become founders. I want to thank you all for taking me along on your ride, and being your partners. You taught me a lot, and I can’t wait to build bigger and better alongside you.
Thank you, architects !
Your biggest and staunchest fan as you become founders.
P.S. special shoutouts to a few of my most favorite archi-founders:
I prefer to start my AEC_VC posts with a “why now” trigger. This week, it would make sense to start with the whole $hitshow that FTX has created.
But my fascination last week was all about a favorite US company – Forge – and how they create a new tech-enabled category of employment in the blue-collar trades and crafts.
So how to connect this to how FTX triggered this post ?
How about: FTX lost a lot of money of hard-working folks in the trades and crafts who invested with them.
Those folks who TRULY keep our economy running, and build stuff.
Let’s talk about them, and how the next-generation of blue-collar 2.0 will look.
In the Western markets, inflation has hit hard over the past 18 months. It peaked around 10% in US and Eurozone in 2022 (with a history of moving the goal-posts).
Why is this relevant to blue-collar ? Because inflation, in its most fundamental essence, happens when there is too much demand for too little supply. The more money is being brought into circulation (M1 supply) and being used frequently (velocity), the more money facilitates demand to bid for supply. If we now have stagnating or – worse – shrinking supply, your inflation becomes painful.
And that’s exactly what happened in the Western markets since beginning of COVID.
Regardless of what the Fed would like to make you believe, inflation in the Western markets is here to stay unless we fix supply.
Because demand is going nowhere. Take housing: The US needs 2.3M new homes per year, Germany and UK 400k and 300k respectively. On top come several hundred thousand of thermal renovations across Europe and US that will be required to deal with the energy situation and climate asks.
While demand will stay, supply of craftsmen and installers is shrinking drastically.
UK lost 175’000 (6%) construction workers just in 2020, and will lose another 500’000 in next 10 years.
Germany needs to fill 190’000 jobs for craftsmen in 2022 already.
US has 2.2M construction workers retiring in next 3 years and will need another 650’000 ON TOP of the 2.2M per year to cope with demand.
Demand rising. Supply shrinking. Inflation here to stay.
Lucky for us, this is where the story begins.
We see insanely talented founders who build the next-gen of blue-collar models to fix supply. Three models in particular create the blue-collar 2.0 space:
Tech-enabled employer: Standardize all workflows. Train lateral/migrant/GenZ hires on core software stack and useful hardware in internal academy. Employ post training. Example: Forge.
Business-in-a-box (BiaB): Enable tech-enabled ownership of craftsmen/installer businesses. Example: Lun.
In new construction, we observe all 3 models and combinations thereof. While in renovation, robotics are much more tricky due to less controlled environment – a combination of tech-enabled employers and BiaB work very well in reno.
We believe the academy piece to be critical to both models. Classical blue-collar education has steep on-ramps for the new talent pools of lateral hires, migrant talents and GenZ+. In Germany (representative for Europe), a carpenter undergoes 1’200 classroom hours over 3 years and has to take 2 exams to be allowed entry into the trade.
There’s so much wrong with this. Monopolistic authority over the curriculum unfit for 2023, too generalistic education (rather than verticalization and further specialization to allow fast entry into productive life), and training for flexible problem-solving rather than being linked to standardized workflows creates the steep on-ramps and too long training time.
This is where tech-enabled employers and BiaB solve the EdTech element of blue-collar 2.0. Having (1) standardized workflows around a core software/hardware stock and (2) deeper verticalized and specialized job roles that they train (3) in internal academies.
Makes for a very sticky blue-collar 2.0 employment. And low on-ramps.
Can’t wait for these next category leaders in blue-collar 2.0 to be built in Europe.
It’s election time in several major economies. One of which is Brazil, which is home to one of the world’s largest biomes – the Amazon rainforest.
One of the presidential candidates in Brazil ran on a “zero de-forestation” campaign. Looks like he won. That’s certainly a step forward.
If you follow my work regularly you know that I’m about the architecture, engineering and construction industry – not the wood industry. It just so happens that A TON of politicians and lobbyists in the Western markets seem to have fallen in love with using a ton of wood in construction recently – other than this Brazilian politician.
Read the primer? Great. Let’s continue part-deux here.
Various scientists come to the conclusion that scaling up the use of wood in construction can significantly reduce CO2 in the atmosphere. Latest estimates point to 30% reduction or more if we use timber buildings for 50% of our new construction.
Here’s the problem that lobbied Western politicians don’t get:
We have 4 billion hectares of forests
They produce 4.75 billion m3 of net incremental wood growth
Of which we can safely harvest 3 – 4 billion m3 per year
In 2020 we consumed 5 billion m3
World population keeps growing by 30%
Middle class keeps growing and emerging economies have a right to consume like their Western counterparts did and do
Therefore we will need to construct +70% floor space in Europe, +90% in Americas, +125% in APAC and +300% in Africa and Middle East until 2060
And the whole point is to use MORE WOOD in construction of those floor spaces.
This means: We are already consuming more than our supply, we are increasing our demand even before we put more wood into construction, then we need to use even more wood in construction, and by doing all this we will kill our wood supply.
Huh.
So maybe some lobbyist will tell some politician over fancy lunch: “We’ll just recycle wood”.
Not happening, amigo. Wood is a downcycling material due to its specific properties. In the US, 17% of wood waste is recycled once or twice. 83% are downcycled to be either a fuel or compost/landfill. You can’t change the shape and structure of wood the way you can with metals/steel.
Meaning: Supply is supply. Not much we can do other than not fuck it.
What we need to fix is the consumption of wood in other sectors to allow construction to use much more wood.
So who consumes wood?
51% for paper & packaging
20% in construction
18% for cooking, fuels, chemicals industry and others
11% for furniture
Ikea alone is consuming 1% of global wood. Furniture + packaging.
So the way I see it, we have amazing founders working on decarbonizing one of the largest CO2-emitting sectors (construction) by bringing new materials and wood into buildings, such as Mighty Buildings, 011h and Juno.
But our politicians need to enable these founders to use more wood by eliminating the use of wood in packaging, paper, fuels, cooking and furniture.
Amazon needs to replace paper-based packaging with new materials.
Ikea needs to replace wood with new materials.
And Western politicians need to equip emerging markets with non-wood-based cooking and heating alternatives, and also need to ban cozy fireplaces for role-model reasons.
Only then we can free up wood supply for scaling up wood in construction, bring construction emissions down to 10% or less of world emissions, and not fuck the (Amazon) forests.
I became quite fascinated by how eco-friendly wood in new construction is. For one, because we partnered with numerous founders over the years who use technology to make the use of wood in buildings more efficient, for example:
But also because I had chats with people from the wood-processing and forestry businesses.
One chat in particular last week stood out. It came with a European family owning a billion-dollar business in wooden construction materials. They asked me flat out: “Where do you see wood in construction going?”
No one is fully immune to media manipulation and public debate. So my first nano-second reaction was “hey, of course wood is great and steel and cement are dirty, so it must grow in use”. But does it? My interest was piqued.
To answer the question where wood in buildings might be going, we have to start by understanding the CO2 footprint of construction and buildings. That’s what I’m doing this week – and will follow up with part two on wood specifically next week.
We all know that the world population continues to rise. 10 billion people are expected by 2060, which is another +30% vs. 2020. And middle classes in emerging countries are rising, too, which means they have more money to spend on consumption. Just from 2009 to 2020, the middle class almost tripled in 11 years.
And that’s the issue with CO2. The demand. Per capita, CO2 has increased by +55% to 4.7 tons per capita compared to 1960.
CO2 is not just a world population issue. It is driven by consumption per capita !
Meat +85% per capita. Cars + 125% per capita. Oceanfreight + 100% per capita compared to 1960. Per person.
And the most dramatic CO2 per capita increase comes from construction. Take cement: +270% vs. 1960 and +135% in just last 22 years.
The issue is per-capita CO2 consumption.
And construction is a massive consumption driver. 17% of the world’s emissions come from construction. Another 17% from operating residential buildings, and 11% from operating commercial buildings. (side note: this is freaking why we need heat pumps across the Western World – go and buy yours with Lun)
However: If you ANNUALIZE the CO2 demand from constructing and operating a building, the design and construction is a 10x higher CO2 driver than the operations phase – because the construction emissions are set free in first 1-2 years, while the ops emissions are set free over the 30-50 following years.
That’s why the 17% of construction vs. the 17% of residential building ops are not the same. The Construction emissions are a 10x larger and instant lever.
I have a feeling people begin to realize this massive CO2 reduction lever that construction is.
However, I fear it is becoming (again) an emotional and non-scientific debate. Much like energy choices after Fukushima and vaccine choices during COVID – experts do not get heard and lobbyists and politicians lead emotional debates.
I fear the same for the use of wood in construction to decarbonize the construction of buildings.
I will write here next week about why wood will make a great contribution to CO2 reduction, but scaling it massively is not going to be a very healthy choice for decarbonization targets.
The same way scaling gas consumption post 2011 was a very bad choice for the EU.
Sources
World Bank 2018 | OECD 2020 | IPCC 2017 | Global Carbon Project 2017 | Oxford 2019 | OICA 2016 | UNCTAD 2018 | Cembureau 2015 | BNP 2018 | UNFCC 2019 | UNFAO 2018 | IEA 2017 | European Commission 2018 | Rocky Mountain Institute 2018 | UN 2017 | Ecoinvent 3.1 2018
But that’s not where the story ends. With the economy in tumble, I wonder if some (if not all) of these players to be active acquirers in 2023 and 2024. Hexagon seems to already lead the charge with 5 acquisitions year to date in 2022.
So I became fascinated by the logical follow-up question: What might happen in AEC M&A in 2023 ? What strategies and targets are available to buyers?
I prefer numbers. The best numbers available to us to work out available strategies and targets is funding data. It should allow us to look at a large share of the AEC-Tech universe (though not all firms). It should be accurate enough, and much better than no numbers.
6‘900 AEC-Tech firms received some form of external investor funding (incl. VC or corporate investments) since 2005 until October 2022. In total they fundraised $120B – no chump change.
Of those 6‘900, 560 have been acquired (not counting IPO‘s). A meager 8% !
Then I pulled a looooooong list of active AEC-Tech firms who have not yet been acquired, not IPO‘d and not deadpooled. This gave me ca. 4’000 firms, the vast majority of which are software. These firms generally could be seen as targets by potential buyers in 2023 and 2024.
I was wondering how a buyer in 2023 might view the landscape. One approach someone in corporate M&A told me is that they use four characteristics to describe AEC-Tech firms.
Is a firm young? (I used 3 years since founding to count as young). A founding team will generally be more willing to sell in year 7 than in year 1.
Has it raised capital recently? (I used the last 2 years). Founders will be more inclined to sell when no new capital sources are available.
Has it raised a lot of capital before? (I tested $5M+ and $10M+). It will somewhat correlate with higher cash burn and thus the need to entertain external offers for financing or M&A when runways shorten.
Does it have exciting/differentiated tech? (A generic product in a competitive market looks more enticing to sell when offers are made). This one is subjective, so I relied mostly on the previous 3 characteristics.
Based on this, the following picture of which firm might be available emerges:
I was particularly fascinated by two clusters of AEC-Tech firms for this debate:
Firms in difficult cash positions (either young and no differentiated tech but high burn, or older firms with or without differentiated tech) in the anticipation they might face a difficult fundraising environments. More a 2023 target. I call this Strategy 1.
or older firms who have raised in the last 2 years, preferably with interesting tech or otherwise good businesses in the anticipation they might reach a plateau in the next year. More a 2024+ target, but opportunities to build a pole-position relationship early might be available in 2023 already. Let’s call this Strategy 2.
For Strategy 1, I was able to find ca. 4’380 AEC-Tech firms right now. Of those, 1’390 who raised last between 2017-2020, so prime candidates for a discussion. Some acquirers might be having debates to get in front of the firms with the most exciting products or ARRs, and find ways to have acquisition discussions. I imagine a big challenge will be to filter effectively, as a lot of these firms will require cash for a reason … but how to know outside in? Protracted discussions with all 1’390+ firms will hardly be possible. Probably only a small fraction of them. Positive is that founders might be more motivated though.
Ca 1’180 firms are available for Strategy 2. 75 firms raised $100M+, 180 firms $20M+, 130 firms $10M+, and 130 firms $5M+. The entire rest raised less than $5M and might not be that interesting. I wonder if an interested buyer would first focus on the middle of the pack with $10-100M raised, since the quality of these firms might be highest. Remember the 8%/560 firms acquired in last 18 years I mentioned at the top of this post? My guess is that a fair share of those 8% in the NEXT years will come from these Strategy 2 targets. Here founders and shareholders might not yet be particularly interested in soliciting M&A offers right now, as they probably see their firm on the up and promising. It might be that interested buyers find more luck in taking minority stakes, building an internal view how the tech and business performs, and moving into pole position to make future offers in 2024+ if the firm begins to plateau.
Strategy 2 could allow for high payout if executed well. The challenge to overcome might be the adverse signaling that founders and shareholders will be afraid of when a potential strategic buyer takes stakes. Something to consider for the acquirers, I imagine.
Overall, it looks like some 400+ potential AEC-Tech firms might attract interest by potential acquirers in the next years. I expect smart buyers to follow one or both of the above strategies. Both are not easy to execute since information is hidden and adverse signaling is real.
My conclusion of last week and this week is: it looks like a lively 2023 in AEC software M&A might be on the horizon.
3 triggers that stimulated me last week to look deeper into M&A in AEC. First, various friends who work at AEC software firms recently debated about M&A going into 2023. Second, Autodesk received another open hate-letter about their (M&A) practices. Third, we’re heading deeper into the 2022 crash – and thou shall be greedy when others are fearful.
I checked out 9 well-known (hybrid) AEC software firms and their M&A track record – including the likes of Trimble, Autodesk, but also game engine devs Unity and Epic to check for alternative strategies.
I went as far back as 2005. My colleague Fabio 👏👏👏 did a helluva job researching and harmonizing the numbers (thanks Fabio !).
331 M&As were done by these players in the last 18 years.
The trend is up – M&A activity almost 2x’ed over this time frame.
2009 was a crash-related down. We already see the same in 2022 – while 2021 was THE record year with 31 deals, 2022 YTD is at half of 2018 (15).
What surprised me was the makeup of those deals. First, by acquirer, and second, by acquiree.
Autodesk might get all the attention (and hate) for their product and M&A strategies.
But it’s hybrid hardware+software players such as Trimble, Hexagon and Faro who dominate M&A. They make up more than 50% of all M&As.
Autodesk is rightly in the discussion, though, making up 20% of all acquisitions (67).
It’s fascinating to observe outside-in that Trimble v. Autodesk seem to follow completely different strategies. While Trimble goes after targets that enable an “on-site” strategy, Autodesk seems more enamored with a “3D-enabled” roadmap.
Hard to say by the numbers how these acquisitions drive a product strategy. Although the numbers give an idea why 3D-users are fearful in their open hate-letters to Autodesk…
But it’s not just behemoths Trimble and Autodesk who we should look at. Over time, M&A priorities clearly shifted. 10 years ago, acquirees used to be much more in hardware and GIS/mapping – those two sectors got hammered.
On the other hand, acquirers today love (1) 3D, gaming/assets, field mgmt and (2) a much more diverse range of (specialty) applications than they used to. 61% of all M&A’s in the last 8 years happened in these spaces.
A special note I found intriguing was that – while 2022 M&A dropped hard by 50% – Hexagon acts contrarian and has their most active M&A year since 2016 (5 deals already).
Yet 2022 deals are down. Multiples (not analyzed, but anecdotally) have come crashing hard, by factor 5-10x or more in some crazy places.
I hypothesize that while seasoned acquirers such as Trimble, Autodesk and Hexagon (each bought more than 60 firms over the last 18 years) are on the prowl for great deals – many shareholders of fantastic software firms are not particularly interested in entertaining shitty deals right now or next year.
So what gets done in 2022 and 2023 ? I wonder if it’s mostly firesales/distressed deals. Or in other words – dogs ?
The numbers I found, plus common sense, don’t suggest that shareholders in great firms in the sought-after categories such as 3D or field management with strong cash positions are particularly keen to sell in the next year.
In my conclusion, acquirers should consider a dual strategy: (1) acquire great products/tech in distressed situations for affordable money (even if it means the multiples look off) and (2) find ways to engage meaningfully with good businesses which still grow in the ballpark of 1.5x-2x YoY in current environment and you expect to plateau on their own in a year or two from now, and then be in pole position to strike a deal at fair multiples in 2023 or 2024.
Looking at our financial markets in early October 2022 is like riding one of these rollercoasters that take you up a loooooong time, only to become really sloooooow before you reach the apex.
You know markets will come down hard. This wait seems like an eternity.
Can we learn to benefit from the uncertainty ahead ?
Nicholas Taleb is a leading economist of our time, who made waves with theories around the 2008 financial crash. He explained black swan events in 2007, and posited that growth in uncertainty flourishes under barbell strategies due to anti-fragility.
For Taleb, anti—fragility is fundamentally about creating optionality and accepting that knowledge is always limited, while exploration under trial & error is unlimited. Taleb says that if trial & error costs less than the upside of being right once – we should adopt an anti-fragile strategy of learning through trial & error to maximize outcomes.
This inspired me to spend to explore anti-fragility further – and look at it in the context of construction.
If we dumb it down, construction has two types of projects: n = 1 projects, where every project is unique in many ways, and n > 1 projects, where we repeat a project serially (or parts of it, for example volumetric modules).
Taleb has identified 7 rules of when anti-fragility applies and an optionality and trial & error maximizes the outcomes – and not seriality.
I’ve analyzed and applied the 7 rules of anti-fragility to both construction project types. You can find the analysis in my slide deck.
To my surprise, it seems to suggest that construction has higher chances of maximal long-term results BECAUSE it widely uses an n = 1 approach. All 7 rules of anti-fragility seem to check out with n = 1 construction projects.
When I interpreted the results, there is a way to make sense of it. Construction is known for relying more on people and their coordination and trouble-shooting than on plans. It’s because construction projects have a lot of interdependencies and externalities. Within the anti-fragility world, top-down derived strategies are inherently less likely to achieve the optimal outcome than a trial & error approach led by the construction project managers and workers on the ground dealing with the interdependencies.
I also find it fascinating in the context of asset-heavy serial modular construction companies, such as Katerra (which went out of business in mid 2021). Anti-fragility suggests that asset-heavy modular contractors will suffer from avoidance of optionality through trial & error which leads to slow knowledge acquisition. Looking from the outside-in, this seems to be consistent with Katerra’s well-documented shortcomings.
It’s fascinating to think that either construction has already found the most efficient learning mode by being n = 1.
Or that projects which are n > 1, asset-light supply chain integrators such as 011h or Juno who facilitate existing competencies in the supply chain and allow for prototyping and testing are more optimal than asset-heavy models.
The EU already had aggressive decarbonization targets for its homes before the war in Ukraine. With inflation ravaging due to money printer go brrr, and war leading to shortage of cheap fossils across Europe – I became fascinated by how European homes are actually heated and insulated going forward.
As it turns out: Europe’s homes are reliant on low-quality energy – fossils. And their insulation is worse than I thought. The most idiotic yet true answer I found are air-gap-walled homes.
A home with an air-gap wall is simple. Built between ca. 1900 and ca. 1980 across European countries – e.g., Germany, Netherlands and the Nordics – it is a home that has a double outer wall. Think of it like a sandwich where the cheese and ham is just air.
The double wall consists of a thicker interior-outer wall, against which a thinner outer wall, or layer of brick clinker, is built. In between the two walls is air. The air is supposed to function as a natural insulator. Spoiler: its insulation is poor.
It is estimated that up to 27 million homes in Germany are constructed with air-gap walls. An additional 10 million homes across Netherlands and the Nordics. On top of any other European countries I did not research.
While that (lack of) insulation is worrisome in light of current events, what really blew my mind is just HOW reliant these home owners are on fossils. 79% of single family homes and 82% of multi-family homes across the EU+UK heat their homes with gas, oil and coal.
And we pay dearly for it. 34 million households were in energy poverty prior to the Ukraine war. In Germany the expectation is that gas bills for home owners will routinely exceed €3000+.
All this when powering and heating homes is already responsible for more than 20% of CO2 emissions.
As it turns out, we have a proven way to insulate air-gap-walled homes – which are heated 80%+ by fossils – with a retrofit cavity insulation. The technology and approach are proven.
A trained craftsman can tell with 95% certainty just by the age of the home and its construction if it is air-gap-insulated. If it is, insulation foam or granule can be injected into the air-gap easily.
The effect: 50%+ heating cost reduction and a 25%+ CO2 emissions reduction. It costs between €5000 and €10’000 for regular homes.
So it has a fantastic ROI.
However, there are only a couple hundred SMB businesses in Germany doing this retrofit cavity insulation. And many home owners won’t be able to pay it in cash and will need financing.
Time to change that. Who’s working on a tech-enabled insulation business, and can scale up cavity insulations?
https://docs.google.com/spreadsheets/d/e/2PACX-1vQUe7iWa59eUAt0KoG0v5uDKsltPsZBjRsVAeFdXvnZs1IuGFdJztsO6TQWSZrz63Whuli50gvTLYFW/pubhtml
http://aecvc.com/
