The "Discovery Trap" in Venture Investing
And why scientific "breakthroughs" rarely capture the value they create.
Science is hard. In the last decade, our partners and members of our technical expert network have led teams that have scaled supercritical CO2 processes, developed novel hypergravity pharmaceutical formulations, and engineered unique sterilization and control system hardware from scratch. We’ve done a lot of science, and we know what it looks like. But investing in science is hard for early-stage technology VCs due to what we call the Discovery Trap.
Discovery Traps have two main characteristics: (1) their core risks generally look like engineering risk on the surface, but in reality are science risk, and (2) when successful, the value captured after taking science risk is often a tiny fraction of the value created.
We break down a bit more on how we think about the engineering vs. science distinction below1:
The return profiles in Discovery Trap investments tend to look like high-margin recurring revenue software returns—capital efficient businesses (in the long run), with recurring revenues (royalties! license fees!), and defensibility (patents)—making them attractive for tech VCs (on the surface…). But in our experience, while the value creation from a scientific breakthrough is often obvious, we find the value capture portion of the equation to not be as simple. As one example, scientific breakthroughs are often followed by significant infrastructure requirements to scale production or a complete rethink of a supply chain of key inputs, both requiring massive amounts of capital to fully capture the value created by a scientific breakthrough. This capital is also often expensive as it comes during the long period between breakthrough at bench scale and product in market at commercial scale (some might call this the “valley of death”), making these businesses very dilution heavy for early-stage equity investors.
Given the difficulty on the value capture side of the equation for scientific breakthroughs, specialist investors exist to shepherd them to market and tactically manage some of the scale-up challenges and capital intensity of commercializing these innovations. Tech VCs like Rhapsody Venture Partners, DCVC, Osage University Partners, and FirstSpark Ventures have developed specialized approaches, teams, partnerships, and skill sets (focused on breakthroughs in areas like AI, Cyber Security, Climate, and Compute, to name a few) that allow them to carefully curate and select founders and technologies that will be able to weather the uncertainty2 posed by scientific development timelines.3 Life science VCs (firms like Flagship, Versant, and Third Rock, to name a few) have similar specializations and unique relationships on how to commercialize novel medical devices and therapies.4
To avoid the Discovery Trap, we spend an inordinate amount of our time developing absolute clarity on what type of risk we are taking (engineering vs. science), how value will be captured in the long-term, and what the capital journey realistically looks like. With alignment on the major technical engineering milestones, we’re able to help companies develop clear value creation milestones and timelines that derisk the development journey and create meaningful design and software intellectual property along the way. This setup makes it easier to capture the value created by the underlying core technology, reduces the cost of capital as a company scales, and creates outputs that a company can defensibly commercialize and sell (specifically, products and services vs. licenses/royalties), often resulting in a monopolistic market position.5
It’s important to state the obvious…the investment opportunities that exist at the intersection of venture-scale markets and engineering-only risk are far and few between. While there are thousands of scientists looking for venture funding every year, there’s far fewer exceptional engineers leaving NVIDIA to start a compute hardware business in that same year (as just one example…). This is why we focus on a few exceptional founders every year that fit the latter description, and when we find them we bet big with high conviction. That’s how one can create the next Apple, SpaceX, or Illumina! If you think this describes you or what you’re working on…reach out!
+Mike
If you’re interested in a deeper dive on this, BCG has a great piece titled “Deep Tech and the Great Wave of Innovation” linked here.
There’s an excellent piece by Jerry Neumann in Reaction Wheel titled “Startups and Uncertainty” where he distinguishes (in great depth) between Risk and Uncertainty. Highly suggest a read!
As with anything, there are of course exceptions. Genentech is one, with Kleiner Perkins making a generational return on their investment. Harvard Business School has a case study titled “Kleiner-Perkins and Genentech: When Venture Capital Met Science” available for purchase if you’d like further reading on it.
We also wanted to acknowledge the role government grants, research universities, and private foundations play in funding the riskiest portion of the development curve, well before private venture funding becomes available. But that’s likely it’s own post entirely!
Illumina, one of the better performing venture investments of all time, is a good example of this. The core technology/science IP was developed in the early/mid-90s, well before the first venture capital dollar was invested in 1998, which set the company on a path to engineer their sequencing device as their primary means of value capture for the underlying scientific innovation.