Foundamental | Defensibility Dilemmas⎟AI Startups, Low-Carbon Cement, And Venture Capital Predictions

Defensibility Dilemmas⎟AI Startups, Low-Carbon Cement, And Venture Capital Predictions

October 11, 2024

We dive into AI startups' defensibility, the challenges of low-carbon cement, and reflect on past VC predictions. Discover insights on investment strategies and industry trends.

Patric Hellermann

Defensibility Dilemmas⎟AI Startups, Low-Carbon Cement, And Venture Capital Predictions

tl;dr

AI startups face challenges in creating defensible positions beyond LLMs

Low-carbon cement innovations struggle with scalability and market adoption

Venture capital allocation in scientific breakthroughs needs more expertise

Overvalued startups from 2018-2021 face uncertain futures

Past VC predictions show mixed results, highlighting industry unpredictability

Robotics investments benefit from software-focused founders

The problem lies with the LPs with the capital allocators, they give their capital to the wrong general partners.

🎧 Listen To This Episode

AI startups face challenges in creating defensible positions beyond LLMs

We're living in a time where AI startups are popping up left and right, but the question on everyone's mind is: how defensible are these businesses? It's a tricky landscape to navigate. Many companies are building on large language models (LLMs), but that alone isn't enough to create a moat.

The real challenge lies in finding ways to differentiate beyond just using LLMs. We're seeing companies try to build partnerships or focus on specific verticals, but it's not always clear if these strategies will pay off in the long run. The pace of innovation in AI is so rapid that what seems cutting-edge today might be commonplace tomorrow.

Take Findable, for example. They've raised a significant round to tackle building documentation management. While they've got some traction and customer love, the question remains: is their technology truly unique, or could it be replicated by others as LLMs become more advanced?

We're noticing a trend where investors are struggling to differentiate between truly innovative AI companies and those that might be more hype than substance. It's a bit like throwing darts at a board – some will hit, but many will miss. The key for founders is to think beyond just the AI technology and consider how they can create lasting value that others can't easily copy.

Low-carbon cement innovations struggle with scalability and market adoption

The quest for sustainable building materials, especially low-carbon cement, is an ongoing challenge in our industry. We've seen countless headlines about breakthroughs, but the reality is much more complex. Take Sublime Systems, which recently secured investment from Holcim. While it's exciting to see big players backing these innovations, we've got to ask some tough questions.

First off, the science behind these materials is incredibly complex. It's not just about creating something that works in a lab – it's about scaling it up to industrial levels. Many of these solutions work great on a small scale but fall apart when you try to produce them in large quantities. The energy requirements or production costs often make them impractical for widespread use.

Then there's the adoption issue. Construction is an industry that's notoriously risk-averse, and for good reason. When you're building structures that need to last for decades, you can't afford to take chances on unproven materials. Architects and engineers are under increasing pressure to personally sign off on the safety and compliance of their designs. In this environment, specifying a new, low-carbon cement that doesn't have a long track record is a tough sell.

We're also seeing a disconnect between the enthusiasm for these materials and the realities of the market. Sure, you might get some high-profile projects willing to pay a premium for low-carbon cement – think tech companies building new campuses who want to tout their green credentials. But for the vast majority of construction projects, cost is still king. Until these new materials can compete on price with traditional cement, widespread adoption will remain a challenge.

The path forward isn't clear-cut. We need continued investment in research and development, but it's crucial that this investment is guided by scientific expertise rather than just throwing money at the problem. There's also a need for more real-world testing and pilot projects to build confidence in these new materials. It's a long game, and we've got to be patient while also pushing for progress.

Venture capital allocation in scientific breakthroughs needs more expertise

We've got a bit of a problem in the world of venture capital, especially when it comes to funding scientific breakthroughs. There's no shortage of money flowing into areas like materials science and building tech, but the way it's being allocated isn't always ideal.

Here's the issue: a lot of the people managing these massive funds don't have backgrounds in science. They're smart folks, don't get us wrong, but they're often coming from business or finance backgrounds. When you're dealing with cutting-edge research in fields like low-carbon cement or advanced materials, that lack of scientific expertise can be a real handicap.

What we're seeing is a tendency to invest based on narratives and charismatic founders rather than the underlying science. It's not entirely their fault – it's incredibly difficult to differentiate between genuine breakthroughs and what might turn out to be vaporware if you don't have the technical background.

This leads to a scattershot approach where capital gets spread too thin across too many projects. Instead of concentrating resources on the most promising scientific advancements, we end up with a lot of underfunded projects that struggle to make real progress.

What we really need is a shift in how this capital is managed. Imagine if we concentrated these funds in the hands of a smaller number of general partners with strong scientific backgrounds. These folks would be better equipped to evaluate the potential of different technologies and could make more informed bets on which projects are likely to succeed.

The other piece of the puzzle is changing the incentive structure. Right now, many fund managers are incentivized to raise and deploy capital quickly rather than focusing on long-term scientific progress. If we could align these incentives more closely with actual breakthroughs and industrialization of new technologies, we might see better outcomes.

It's a tricky balance to strike. We want to encourage innovation and risk-taking, but we also need to ensure that capital is being deployed effectively. For founders working on genuinely groundbreaking technologies, this current environment can be frustrating. They might find themselves competing for attention with flashier but less substantive projects.

The bottom line is that we need a more nuanced, scientifically-informed approach to funding breakthrough technologies in our industry. It's not just about throwing money at problems – it's about making sure that money is guided by real expertise and focused on the projects with the best chance of making a tangible impact.

Overvalued startups from 2018-2021 face uncertain futures

We're looking at a bit of a pickle for a bunch of startups that raised big money between 2018 and 2021. Here's the deal: during that time, venture capital was flowing like crazy. Companies were raising huge rounds at sky-high valuations, often with very little revenue to show for it.

Fast forward to today, and we've got a lot of these companies sitting on decent revenue – we're talking $50 to $200 million – but they're in a tough spot. Why? Because their last funding round valued them way higher than what they're actually worth now.

Let's break it down with an example. Imagine a company that raised $100 million at a $500 million valuation back in 2021 when they had zero revenue. Now, they might be pulling in $10 million in annual recurring revenue. That's not bad, but it's nowhere near enough to justify that $500 million valuation in today's market.

So what are their options? Going public is off the table – they're not big enough for an IPO. Acquisition is tricky because who's going to pay that inflated price? And raising more money means facing a potential down round, which nobody likes.

This isn't just a problem for the companies themselves. It's causing headaches for their investors too. VCs are sitting on paper valuations that don't match reality, which makes it hard for them to report accurate returns to their own investors.

In the construction tech world, we've seen less of this craziness than in some other sectors, but it's still out there. The companies that are in the best position are those that have managed to get close to cash flow break-even. They might not be the unicorns everyone dreamed of, but they've got a shot at building sustainable businesses.

For the rest, it's going to be a bumpy ride. We might see some tough negotiations in the coming months and years as these companies try to figure out their next moves. Some might manage to grow into their valuations, but others are going to face some hard realities.

The lesson here? Valuation isn't everything. Building a solid business with real revenue and a path to profitability matters more than a flashy number on a term sheet. As we look at investing in new startups, we're keeping this lesson front and center.

Past VC predictions show mixed results, highlighting industry unpredictability

We've been taking a look back at some venture capital predictions from 2018, and it's a bit of a mixed bag. It's always fun to play the prediction game, but as we've seen, getting it right isn't easy – even for the pros.

Let's start with what they got right. The prediction about bigger seed rounds? Spot on. We've definitely seen that trend continue. And the call on AR needing more time to develop? Yep, that was accurate too. We're still waiting for AR to really break into the mainstream in a big way.

But then there were some misses. The idea that there'd be more liquidity through IPOs and M&A? Well, 2021 was a banner year for that, but it wasn't a consistent trend from 2018 onwards. And the prediction about blockchain and ICOs finally delivering a product and being disappointing? That's a complicated one – the crypto world has been on quite a roller coaster since then.

One interesting prediction was about increased investment in frontier tech – things like autonomous vehicles, robotics, and other hardware-heavy innovations. While the timing might have been a bit off, we are seeing a surge in these areas now. It's just happening a bit later than predicted.

Here's a fun one: the prediction that Apple and Microsoft would get into a bidding war over Magic Leap. That definitely didn't happen. It's a good reminder that even when something seems like a sure bet, the tech world can surprise you.

What can we learn from all this? First off, making predictions in the VC world is tough. The pace of change is so rapid, and there are so many factors at play. It's easy to get caught up in hype cycles or miss the next big thing.

For us in the construction tech space, it's a reminder to stay flexible. We can make educated guesses about where the industry is heading, but we need to be ready to adapt when things don't go as expected. It's also a good lesson in humility – even the smartest folks in the room can get it wrong sometimes.

Looking ahead, we're trying to balance being forward-thinking with staying grounded in reality. We're excited about the potential for AI, robotics, and sustainable materials in construction, but we're also aware that progress often happens more slowly than we'd like.

The key is to keep our eyes open, stay curious, and be willing to change course when new information comes in. That's how we'll navigate the unpredictable waters of construction tech and venture capital in the years to come.

Robotics investments benefit from software-focused founders

We're seeing an interesting trend in the robotics space, especially when it comes to construction tech. It turns out that some of the most promising robotics companies aren't being founded by traditional automation experts. Instead, they're coming from folks with deep software and product backgrounds.

Now, you might think that building robots is all about the hardware – the nuts and bolts, the servos and actuators. But in today's world, the real magic often happens in the code. That's why we're getting excited about founders who come at robotics from a software angle.

Here's why this approach makes sense, especially in construction:

Flexibility: Construction sites are complex, ever-changing environments. Software-focused founders tend to build more adaptable systems that can handle this variability better than rigid, fully automated solutions.
Iterative Development: Software folks are used to rapid iteration and continuous improvement. This mindset is crucial when you're developing robots that need to learn and adapt quickly.
Focus on ROI: Instead of trying to automate everything at once, software-minded founders often look for the specific tasks or processes that will deliver the most bang for the buck. This targeted approach can lead to faster adoption and clearer value propositions.
Integration Savvy: In construction, any new technology needs to play nice with existing systems and workflows. Software experts are often better equipped to handle these integration challenges.
Data-Driven Decision Making: Software backgrounds often come with a strong emphasis on data analysis. This can be crucial for optimizing robot performance and demonstrating value to customers.

We're seeing this play out with companies like Monumental, where the founders' software expertise is a key driver of their robotics innovation. It's not just about building a machine – it's about creating an intelligent system that can learn, adapt, and deliver real value on the job site.

This trend doesn't mean that hardware expertise isn't important. Far from it. But we're finding that the most successful robotics companies in construction are those that can blend software smarts with hardware know-how. It's about finding that sweet spot where code meets the physical world.

For investors like us, this shift means we're looking beyond just the robotics credentials when evaluating founders. We're asking questions about their software development experience, their approach to product management, and their vision for how AI and machine learning can enhance robotics performance.

The future of construction robotics isn't just about bigger, stronger, or faster machines. It's about smarter, more adaptable systems that can tackle the real-world challenges of building sites. And more often than not, it's the software-savvy founders who are leading the charge in this exciting space.