Last week Y Combinator, the well-known and very successful technology startup incubator, announced that it was going to begin experimenting with biotech startups. This announcement came as a surprise to many. Several good posts on the subject appeared after a Nature News piece, including Derek Lowe’s In The Pipeline blog and the Curious Wavefunction blog.
Here’s the gist of the concept: a biotech startup selected by Y Combinator would get ~$120K in seed funding, and lots of advice, guidance, coaching via the “incubator” experience in exchange for some of the founding equity of the startup.
Let me start by saying I think this is a great initiative, and Y Combinator is right to draw the parallels between the capital efficiency trends in biotech and those in software.
In a world where the need for hardware, servers, and computer infrastructure has been replaced by the cloud, mobile, and social media, the cost to launch a tech startup has decreased enormously over the past decade. As my tech partners like to say, startups can do on $500K today what took $5-10M a decade ago.
Although the math may be different, virtual biotechs doing drug discovery today are leveraging a similar trend: remove the heavy fixed costs of building out your own laboratory, purchasing expensive lab equipment, and then having to “feed” the system, and move to a lower cost virtual model of renting lab capabilities via a global network of CROs and collaborators. Others have already commented about the decreasing cost of DNA sequencing (here), but same holds for other aspects of drug discovery, like computer-aided drug design and structural biology. It’s easier to start a scientifically credible biotech today than ever before, and entrepreneurs can make real progress in validating a thesis on seed capital.
But aside from that general theme, which resonates with how we think about many early stage biotech opportunities, I was struck by the criticisms and/or biases about the biotech vs tech narrative around this announcement. It’s not unexpected given the way anecdotes rather than data inform most of the “startup psyche” as well as the Pharma mindset. Two assertions in particular leaped out at me, especially since they are wrong:
- Biotech startups are “riskier” than tech. The risk of failure for a tech startup is way higher than biotech – some 75%+ of tech startups fail historically (lose money for their investors), whereas only 40-50% of so of biotech startups suffer that fate (here, here for more). And, at the aggregate level, their capital adjusted loss ratios have historically been higher too (losing more money when they fail). Its true that drugs face attrition risks of all sorts, but tech companies face a ton of risks too; biotech’s scientific risks are no bigger or more menacing than all the technology and business risks a new software firm faces.
- Biotech is “too expensive” and takes “too long”. Too expensive and too long are relative terms – and relative to what? It may indeed take a long time to go from test tube to commercial sales to patients, but no investor backs a drug discovery play with that entire cycle time in mind; we invest over 5-8 year time frames and anticipate a liquidity event in that window, if not before. Building a big tech company, especially in the enterprise arena, also costs a ton of money and takes years. Two massive Y-Combinator successes, AirBnB and Dropbox, have raised $750 and $1B, respectively. That’s obviously more capital than most 7-year old biotech companies ever raise. The point isn’t how much money you need to develop a drug or a software product, it is whether you can raise the funding at an acceptable and attractive price (which depends on the cost-of-capital, more on that later). It’s also worth noting that the typical (median) private capital raised before Tech IPOs is over $75M (here), not wildly different than biotech IPOs. Furthermore, despite the widely-held view of biotech taking far longer to get liquidity than tech deals, data suggest that historically the opposite is true: biotech startups are often faster than tech startups to get from founding to an IPO or significant M&A event (here for data).
So the data around biotech and tech startups don’t support the above two concerns, and they really aren’t credible criticisms of Y Combinator’s push to consider biotech.
But one relevant criticism is what can really you do in biotech on $120K. It’s a fair point. If the biotech is based on an academic lab’s work, there’s likely some key value-enhancing and confidence-strengthening “validation” work to do around confirming the original findings. Some of this can be done on $120K, such as a couple good non-GLP preclinical models and some PK work. Most of the time it requires 3-10x that amount at minimum to do a “confirmation & validation” process rigorously, but just as a similarly scaled Phase 1 SBIR can help, I suspect YC’s $120K could be catalytic for other seed-stage capital. However, it is certainly true that the seed funding “milestones” are very different in biotech and tech. Measuring weekly growth rates in biotech is not going to work (here).
There are many other fundamental differences between the sectors that impact how one should think about seed-stage investing in biotech, but there is one critical difference worth further comment here as its relevant to this Y Combinator experiment: the rate at which the cost-of-capital can change over time early in the life of a startup.
Tech startups can raise a $500K seed round and a year later, if they “crush it” like everyone does in Silicon Valley, they can go on to raise a $5M Series A at valuations 10x+ higher than the seed round. For a deal with escape velocity, the Series B can be a far higher price than that in another year or so. This doesn’t happen often, but for big tech winners it’s in the realm of the possible. Of the 500+ startups that Y Combinator has backed since 2007, apparently ~10% have valuations (or been acquired for) above $40M (here). Several of those well into the billions of dollars (like Dropbox, AirBnB). So when it happens it represents a dramatic reduction in the cost of capital over the first few years of a startup’s life. Raising more money gets cheaper and cheaper at a rapid rate in the ~10% of cases when things work. By comparison, this rapid valuation trajectory – in a few years, and only a small amount of invested capital, hitting valuations in the $100s of millions – happens with a <1% frequency in biotech, and I’d wager this trajectory won’t happen on only Y Combinator’s $120K invested into drug discovery stage biotechs.
Why does biotech not have this sort of early stage “cost-of-capital change” that some tech winners can achieve? It’s not necessarily about revenues or profits; many high profile tech stories hemorrhage cash for years and postpone “monetization” as they build their platforms. Fundamentally, it’s about supply and demand around great ideas and capital. Investors have to fight to get into hot deals in the tech world, and given the high rate of failure there’s more capital around than ideas likely to work. Seeds can hit the aforementioned escape velocity fast, and understanding the product, market, and talent signals from seed investing can help select high growth winners from others. Because of Y Combinator’s track record of finding and distilling seed-stage signal, many VCs fight to get into Y Combinator tech startups’ seed and Series A rounds. And the diversity of “VCs” interested in investing is huge: super-angels, micro-VCs, established brands, mega-funds, upstarts, etc… Furthermore, it’s not in a tech entrepreneur’s interest to lock up multi-year funding (whether tranched or not) if the cost-of-capital has any meaningful potential to continue to improve quickly (i.e., valuations likely to continue to go up over next 12 months). Data flow is instantaneous in many tech settings, so tracking progress is straightforward and leads to metric-driven valuation models. In a market with a reasonable and liquid supply of growth capital, a startup would much rather “test price” for the Series B in 12 months rather than locking in a price and taking the dilution of a bigger round now. This sense of both scarcity and relative capital provider abundance leads to early stage tech startups’ impressive pricing power when they “work”, a premise that rests at the heart of “big hit, low hit-rate” early stage tech investing.
Biotech is vastly different. While we have clear signals we look for during the seed phase too – for instance, derisking the biomedical science (e.g., replicating and generalizing a founding academics work) – these signals are not enough to create an auction-like fundraising environment for several reasons. First, there are less than a couple dozen active early stage biotech investors, and only a fraction of those actually consider seeding drug discovery startups. Second, there are few if any “metrics” to validate a valuation or a step-up. Data flow in biotech is not real-time or instantaneous; generally, its punctuated by a series of important “events” or study readouts, each incrementally derisking a story. This doesn’t aid in projecting valuations very effectively, and DCF models are useless. Because of this, raising new rounds at higher prices every 9-12 months or so is very challenging. In light of this, having “dry powder” around the table through multi-year often tranched financings is critical in a fickle, cyclic funding market. Locking in an early syndicate that can “go the distance” through the key “events” with a startup greatly reduces the future financing risk and management distraction. Even in the current environment, early round prices or deal structures aren’t going to escalate any time soon in the biotech startup world. While later stage pricing has gone up with the opening of the public markets, it’s still got a long way to go before meaningful exuberance works its way back into seed and Series A rounds (other than the few monster Series A’s being done today). And today’s exit valuations will need to hold longer term for big step-ups to “work well” for both early and late stage investors. Because of this dynamic, seed-stage biotech investing is about tight go/no-go stringency around the critical assets/hypotheses, strategic reserves allocations and equity capital titration, and being able to “go long(er)” on biomedical science that’s working to get to the big “data” events. Big outcomes from early stage biotech investing can and do happen, but these are less driven by halo-deal “hits” and more about steady top decile return potential.
This temporal difference in the “rate of change” in the cost of capital for startups, and its impact on the venture funding cycle, will dramatically impact Y Combinator’s ability to leverage its traditional model with their seed investments in biotech. A key to their success will be facilitated syndicate formation at the time of the YC startup selection, with longer-term capital plans in place to build and grow those biotechs through data events rather than growth metrics; going alone into a drug discovery seed expecting an exuberant private funding market to appear and carry the project forward at attractive prices for the incubator model is unlikely to work, and hasn’t for biotech incubators historically.
I’m certainly not suggesting that’s what Y Combinator is thinking with its biotech efforts, or that they will dive headfirst into drug discovery vs more IT-led life science enabling technologies, but felt given the news this perspective was worth emphasizing.
When done well, seed-stage therapeutic biotech investing works by delivering solid companies with great drugs, and what follows are great returns. The number of notable early stage successes in the recent M&A and IPO window are testimony to this. I am hopeful that Y Combinator’s experiments in biotech deliver more of those successes.
