Busting deep tech myths with Bison Ventures’ Caleb Appleton

Bison Ventures invests in startups building in techbio, climate and sustainability, and AI in the physical world. How did you land on these three investment areas?

They reflect an acceleration we’re seeing at the intersection of physical-world challenges and emergent engineering and technology capabilities. You’re seeing this convergence of new tools, data, and computation engineering that drives drastically more capital-efficient and drastically faster product iteration and learning cycles. Gone are the days of having a single hypothesis tested and learned. We are now in an era of learning programmatically and finding global maximums and minimums much more quickly within a specific domain. And when we looked at our team’s specific areas of investing and technical expertise and where that overlapped with where we see the opportunity in the future, that’s how the three categories shook out.

How does Bison Ventures do technical diligence on startups? 

Everyone on the team has an engineering degree and has been investing across these domains for a decade or more. We’ve all developed a good initial sense of what a Bison portfolio company is, and in specific areas where we have individual domain expertise, we’re able to go much deeper. We also spend quite a bit of time leveraging our extensive network of external experts to the fund in this diligence process. Everything from academics to individuals in industry and other entrepreneurs. We use people who have formal relationships with Bison and those who don’t.

The best technology innovations, by nature, have discourse around them. As an investor, you are looking for experts to tell you that a company is not impossible and that if successful, it will meaningfully move the field forward. You’re looking to deeply understand the critical risks with delivering on the specific startup’s technology vision so that you’re entering this relationship eyes wide open to the challenges. You want to make sure there’s enough latitude so that as both the technology and market evolve, the company has the flexibility to pivot.

And of course, technical diligence is just a piece of the puzzle. It is necessary but insufficient to have a good technology or a good vision; there must also be the right team for the opportunity that the technology unlocks. 

What is your typical check size? 

Our average check size varies based on the specific company and round. We are investing in both Seed and Series A companies. Our first fund is a $135M vehicle, and we expect to make approximately 15 investments out of the fund. That means, over the lifetime of the fund, we expect the “average” company to receive $8-10M with approximately 50% of that coming in the first investment. 

We believe in the power of smart syndication to ensure companies have access to the capital necessary to drive both technical and commercial risk retirement. As such, we lead investments and co-lead with other similar or even larger funds. 

What is the biggest misconception about building or investing in deep tech? 

There’s often a misconception that deep tech companies are “moonshots” by nature. As a result, many capital sources have historically shied away from backing businesses in the category because many view it as “too risky.” Our point of view at Bison is that deep tech startups can be less risky than software businesses, where the idea of product-market-fit is unknown from the beginning. 

For this to be true, you need to be both disciplined about when you enter a company and the types of businesses you back. The point where we invest is typically beyond scientific or hypothesis risk and more in the domain of engineering risk. This might mean a founder has proven a technology can work, but they haven’t shown it to be possible at scale. Or it might be moving from several point solutions to an integrated system that has yet to be proven.

But deep tech is not a binary technology risk that requires hundreds of millions of dollars of capital to understand whether or not it works. There are many businesses where it is a convergence of innovation across several fields that drives this exponential improvement in efficiency. I have the honor of backing a company called Vivodyne that creates clinical-level human data without the need for clinical trials through their HIVE platform, which enables high-throughput production of human representative tissues for the first time ever. Vivodyne is built on innovation in bioengineering, advancements in fluidics and robotics, and finally AI to drive experiment planning and interpretation. As a result, they’ve built a device that enables them to learn 1000x faster than others in the field. 

None of those by themselves are necessarily at the far reaches of their individual domain, but when coupled together, they drive something that’s extremely powerful and extremely hard to recreate. It’s a great example of the type of business we back at Bison Ventures.

If not technical risk, what is the most common reason deep tech startups fail? 

We very rarely see businesses fail because they didn’t deliver on the technical vision. We do see businesses fail because they don’t understand the right business model or how to sell their innovation in a way that drives rapid growth. As a result, we spend a lot of time validating the here and now product-market-fit. For many of our companies, we have a good sense of the market opportunity and size from the outset. We are able to de-risk the commercial path from the get-go by speaking directly to customers and buyers, who can tell us if the business is working to solve critical problems. 

We also recognize that capital access is a critical risk for deep tech businesses because many investors have historically avoided the category. As such, we try to ensure that the round we are backing is the last round where any investor has to believe in a pure technology story, which opens up capital access because investors can start valuing the company on the fundamentals of their commercial business.

Why is there a persistent misconception among investors that building a software company is less risky than building a deep tech company?

One reason why software businesses quickly ‘ate the world’ was the concept of the lean startup and an agile deployment cycle for rapidly testing, validating, and learning from what was being built. The misconception is that a similar cycle is impossible in deep tech. I think we’re in a moment where that’s becoming increasingly possible in deep tech due to a convergence of better supply chains, software, and enabling toolkits from core AI models to manufacturing techniques.

As an example, we have a robotics business in our portfolio named Cobot, founded by Brad Porter, who was previously chief technical officer at Scale AI and head of robotics at Amazon. Unlike many other robotics businesses that we’ve looked at recently, he raised capital in a seed round and within 12 to 14 months had robots deployed at customer sites. This would have been impossible less than a decade ago, when getting to deployment required years and significant amounts of capital. Importantly, Cobot also built its robotic system in a way that is modular and enables rapid iteration without requiring a complete architecture. Now, Cobot is working with significantly more customers, learns from each of those deployments, and rapidly reflects those learnings back in the next product iteration—how’s that for an agile deployment cycle? 

We get really excited about businesses where the founder is incredibly customer centric and building a platform or technology that is flexible enough to meet the market where it is, rather than myopically following their individual vision for where the technology needs to go.

In some cases, founders don’t have the luxury of being able to deploy early and getting customer feedback. In biotech, for example, you don’t know for 7 to 10 years if you’re going to have an approved drug. That said, founders can build partnerships with pharma early on and clearly understand the right product to build through conversations with key opinion leaders and clinicians.

What qualities do you look for in founders? 

First and foremost, we look for someone who has a deep grasp of the technology. This is technical work, and it requires technical individuals to build. It also requires someone who’s charismatic enough to be able to attract capital. A large part of being an entrepreneur is being a salesperson, whether you are pitching to customers or investors. In deep tech specifically, these companies consume real capital and so need to be able to access it now and in the future. 

The second is a holistic team. Very rarely do all of the attributes come within a single individual. We’re looking for the right mix of someone who deeply understands the technology and has a clear vision for how that plays out, and someone who deeply understands the market. Sometimes that’s the same person, sometimes it’s a small team with overlapping but distinct skillsets.

Beyond that, you want to make sure that the right experience from the direct industry they are selling into exists both on the founding team and through advisory networks. It is really hard to build a business; you want people who have been there before around the table to share wisdom and guidance. Ideally, your mistakes (and you will make them!) are novel mistakes. Finally, the question I ultimately always ask myself before investing in a company is whether this is a CEO that I would be very excited to report to and work with every day.

How do you define deep tech?

Deep tech is the convergence of new technologies and large existing markets where those new technologies can drive solutions that are 10 times better, faster, cheaper, and more efficient.