Funding RealTech

Steve Crossan
6 min readJan 23, 2024
An early generated image of the biosphere © DayhoffLabs

TLDR: as we move into the era of RealTech classic venture capital patterns may need to be adapted.

For the last 50 years or so we have tended to fund innovation in society through venture capital. Classically, VC is based on the idea of high risk, small but real shot of high reward.

A venture investor may be happy to invest $1M into a business valued at $10M on day one without revenue or even a settled business plan, based on the idea that an exceptional team has a narrow but real path to grow that business to 20–100x that valuation within only 5–10 years.

Within a portfolio, a few businesses will successfully progress through the rounds based on their performance, selling perhaps 15% of the stock at Seed, 20% at Series A two years later and another 20% at Series B two-to-three years after that so that founders and team are left with about half the business at (say) a valuation somewhere over $100M.

Importantly this is based on very fast growth in revenue and/or usage — investors talk about benchmarks such as “triple triple double double” meaning tripling revenue in each of the 2 years following first going to market, then doubling it twice more. That’s very hard but not impossible for software companies — because for good software delivered online the marginal cost of production, distribution and customer acquistion approach zero. So if you hit a valuable niche with a great must-have product that’s priced well, there are few barriers to taking off extremely fast.

There are limits to the set of problems you can solve in the world with SaaS. Venture investors are increasingly looking at new sectors that combine a software element with building in the real world in some way (‘deep tech’, climate tech, TechBio, software-enabled businesses — let’s call them RealTech). And many of these sectors are where we most urgently need innovation.

But the scaling characteristics of these businesses are very very different. It is hard to triple your revenue in a year if you have to triple the capacity of your factory. And the economics, capital requirements and funding sources for building a plant are also very different.

So what are the possible playbooks for RealTech founders ? Here are a few I think about:

  1. Only build what you can automate. The simplest version of this is only build software but there are some interesting approaches starting to take automation and software-controlled manufacturing into the world of atoms. Anything you can’t entirely automate — find someone else in the value chain to do that bit. One objection to this model in (for example) TechBio is that you leave a lot of the value on the table. You are creating IP which in turn helps create downstream IP, but it’s hard to capture a lot of that if someone else is doing the hard (expensive) yards. Hence lots of TechBios for example end up turning themselves into BioTechs and becoming companies that are valued on their assets not their software — but I would argue they then become a different kind of business. A very small number of companies (Codexis?) have threaded this needle. But if you believe in TechBio¹ (or another TechReal sector) as a founder or investor then you need to articulate that path to a large business with software scaling characteristics.
  2. Change the economics (but not the control). BioTech has long understood this. BioTechs involve labs, reagents, capital equipment, research risk (unpredictable time horizons), and making a bet on at most a handful of programs hoping at least one makes it as a medicine. So BioTech investors typically demand 80% of the business over a few rounds rather than around 50% — lower valuations making it easier to get the growth they are looking for. Non-dilutive funding can be a bigger part of the picture. This model doesn’t sit well with tech investors or founders but it may be more rational for RealTech. My hunch is founders care more about control than maximizing economics. Maybe there are ways to give founders meaningful control for longer, even after they no longer have more than 50% of the economic rights.
  3. Go very big early (e.g. NorthVolt). Do an end run around your scaling challenges by building the biggest version of your infrastructure first. There are some domains where this makes sense — like batteries for example where we know the demand is going to grow very large and there are geopolitical considerations meaning we can also crowd in government funding, forward purchase agreements etc. Tech founders (and tech investors) don’t always know where to access the kinds of funds and expertise required to pull off that kind of scale though (a big opportunity for those that do). Vargas of Sweden seem to be pursuing this investment strategy, sometimes in partnership with the European Investment Bank.

There are also implications for investment models:

  • A bigger role for philanthropy, government and universities (and perhaps corporates) in de-risking the early stages (pre-spinout). In the post-war years a lot more innovation was done this way via e.g. the National Labs in the US (not to mention the role of the SBIC in the early development of venture). There are hints of a revived version of this in the Focused Research Organisation model (ARIA, Convergent Research) and in innovative nonprofits like the Institute for Carbon Management at UCLA — but the scale and impact could be much much larger.
  • I think there’s an opportunity for early stage RealTech funds that look a bit more like BioTech. These would look for more share ownership while finding ways (share classes, covenants) to maintain founder influence a bit longer. They would perhaps have longer time horizons, or evergreen funds. And they’d plug into a network of experienced talent for the infrastructure-building stage. These funds would need to be large enough to support a portfolio of businesses through several stages.
  • First-of-a-kind (FOAK) capital investing needs to mature. Project finance is an established sector, but we need a sub-sector that is willing to take higher risk on newer technologies, again through a combination of portfolio effects and demanding more ownership or seniority. Especially for strategy 3) above we need tested models for and expertise in building big early, which will combine private institutional and strategic capital, government investment, debt, non-dilutive and forward purchase agreements. We are seeing hints of what the government side could do in areas like quantum, but there is scope for much more.

The good news is there are many pools of capital that want exposure precisely to this sector. VC as we’ve known it in the last 20 years will need to adapt to meet that opportunity, changing ownership goals, fund size, portfolio strategy and especially novel blends of capital. Governments have a huge role to play in developing this ecosystem, and the rewards (both financial and for society) are uncapped.

Thank you to Jason Pontin Sam Endecott Lina Wenner Will Wells for reading and commenting and to Clayton Rabideau for conversations that got me thinking about this. Together with Josh Goldford I’m the founder of Dayhoff Labs building foundational models of the chemistry of life to solve cell-free synthetic biology.

[1] By TechBio I mean a company that is using a software innovation to drive innovation in biology giving it (supposedly) software scaling characteristics. By BioTech I mean an early stage science-driven therapeutic company driven by expertise in a particular biology, valued on the therapeutic assets it can generate. By analogy I’m coining TechReal to mean a company that is using software to drive innovation in any Real (atoms not bits) industry.



Steve Crossan

Research, investing & advising inc in AI & Deep Tech. Before: Product @ DeepMind. Founded Google Cultural Institute; Product @ Gmail, Maps, Search. Speak2Tweet.