The Darwin Fund
“Venture capital? Risky!” This is the reaction I often hear when I mention my work. I have come to expect the widening of eyes, a look of concern, and the inevitable question about how I sleep at night knowing that “90% of startups fail.” (FYI, it’s always an accountant or a below C-Suite level exec that think like this.)
This common view of venture capital as extraordinarily risky overlooks a key truth about innovation. The so-called risk is simply the price of exploring the unknown, and without that exploration we would not have Telephones, the Fords, Amazons, or Teslas that have reshaped our world.
Many venture-backed companies do not succeed as originally envisioned. Calling this “failure” is like calling evolution wasteful simply because most species that have ever existed are now extinct. Both are remarkably efficient systems for discovering what works in complex, uncertain environments.
When the founders of Odeo realized their podcasting platform was doomed after Apple introduced iTunes podcast integration in 2005, they did not pack up and quit. Instead, they shifted focus to a side project: a simple messaging service that eventually became Twitter, a company now worth billions. This pattern of adaptation is not unique to startups; it is a fundamental aspect of how innovation unfolds.
Natural Selection in the Wild and in Markets
Evolution does not create perfect organisms. It produces creatures that are good enough to survive and reproduce in their current environment. Over 99% of all species that have ever existed are now extinct, not because they failed, but because they were part of nature’s grand experiment.
Similarly, the startup ecosystem is not designed to yield perfect businesses from day one. It is an experimental system in which many ventures will not survive in their original form, yet the insights, talent, and technologies they develop later recombine in unexpected ways to create extraordinary value.
Consider these evolutionary patterns:
• Adaptation through pivots: PayPal began as cryptography software before transforming into a payment processor. Slack emerged from the remnants of a failed gaming company. Instagram started as a location-based service called Burbn before focusing solely on photo sharing. In one case, 12 out of 13 original features were dropped, much like how slight genetic tweaks in Hox genes can create new body plans.
• Recombination of talent: The so-called “PayPal Mafia” went on to create Tesla, LinkedIn, YouTube, Yelp, and many other groundbreaking companies. This was not coincidence but the recombination of PayPal’s “genetic material” in new environments. Research shows that 38% of second-time founders recruit engineers from previous failed startups, a form of hybrid vigor driven by combinatorial innovation.
• Selection pressure: During the 2008 financial crisis and the 2020 pandemic, startups that could not adapt quickly disappeared, while nimble ones not only survived but thrived. For example, food delivery startups experienced 300% user growth during the 2020 pandemic, whereas brick-and-mortar retail ventures saw a 78% mortality rate. This mirrors mass extinction events in nature that clear the way for rapid diversification.
The Misunderstood Economics of Startup Failure
The oft-quoted “90% failure rate” (1-in-10) in venture capital - which was more accurate 10 to 20 years ago than it is today, leads many to see the industry as fundamentally broken. Critics miss that innovation follows power law distributions rather than normal ones. In a portfolio of 100 startups, one blockbuster success can multiply the entire fund several times over.
This is not a flaw but a mathematically inherent part of exploring unknown territory. As investor Paul Graham once noted, “If 90% of everything is crap, and you want to find the 10% that isn’t, you have to generate and examine a lot of candidates.”
Data supports this counterintuitive reality. A study by Correlation Ventures that analyzed 21,000 startups found that 65% of venture investments returned less than the original capital, while just 4% delivered returns of 10 times or more. Those rare successes generated the vast majority of industry profits. This power law dynamic, where 0.5% of startups account for 60% of venture returns, is similar to how a single mutation in invasive species can reshape entire ecosystems.
What appears as risk to outsiders is actually a carefully designed system for finding those rare companies capable of generating 10 to 1000 times returns. These breakthroughs would not be possible without the willingness to place many bets, fully aware that not all will succeed.
The Adaptive Value of Entrepreneurial Failure
First-time founders have roughly an 18% chance of success. Founders who have failed before see that chance rise to 20%, and those who have previously succeeded can reach a 30% success rate. Why does previous failure correlate with increased success?
The answer lies in adaptation:
1. Tacit learning: Failed founders acquire hands-on knowledge about customer development, team building, and operational challenges that books or classes cannot teach.
2. Network resilience: Entrepreneurs who have weathered failure often maintain valuable relationships with investors, mentors, and talent that strengthen their next venture.
3. Selection effects: The willingness to start over after a setback shows the persistence that entrepreneurship demands.
Modern data paints a more nuanced picture than the “90% failure” myth:
• Year 1: An 18% failure rate for first-time founders compared to 12% for serial entrepreneurs.
• Year 5: A 50% cumulative mortality rate across all startups versus 70% for traditional small and medium enterprises.
• VC-backed ventures: A total failure rate of 25-30% compared to 40% for bootstrapped companies.
These statistics reflect evolved adaptive strategies. The 20% improvement for second-time founders is reminiscent of the Red Queen Hypothesis, which posits that continuous adaptation is necessary for survival.
As Reid Hoffman observed, “The entrepreneurial journey is not about a single expedition; it’s about developing the capacity to lead multiple expeditions.”
Building an Evolutionary Investment Strategy
Viewing startups as evolving entities calls for a fresh approach to investment:
1. Embrace Intelligent Failure
Not all failures are equal. Intelligent failures test meaningful hypotheses about markets, technologies, or business models, while unintelligent failures stem from poor execution, team dysfunction, or flawed assumptions that could have been tested more cost-effectively. Analysis of 500 failed startups shows that 62% had tested value propositions later successfully commercialized by others. Smart investors do not avoid failure; they help companies fail intelligently by:
• Allocating 60-70 percent of capital for follow-on investments in winners.
• Tracking the specific hypotheses each company tests.
• Facilitating knowledge sharing among companies exploring similar ideas.
2. Invest in Adaptive Founders, Not Rigid Business Plans
The most successful founders are not those who start with a flawless plan but those who iterate quickly when reality strikes. Look for evidence of previous adaptation in a founder’s history, ask for examples of times they completely changed course when new information emerged, and value diverse founding teams that can address various challenges. Data shows that 73% of unicorn startups have founding teams that combine technical and business expertise, compared to just 41% in failed ventures. Startups that undergo 3-5 major pivots have four times higher survival odds.
3. Design for Variation and Selection
Just as genetic diversity strengthens species, a diversified portfolio enhances returns. Diversify across business models, sectors, and founder types. Reserve around 15% for contrarian bets that challenge current assumptions, and structure investments to allow for rapid iteration and testing.
4. Facilitate Recombination
The power of evolution comes from the recombination of genetic material. In the startup world, this means creating avenues for talent to flow between portfolio companies, hosting events where founders exchange ideas, and finding opportunities to combine different technologies or approaches. Startups that share APIs or data pipelines have a 22% higher collective survival rate.
5. Accept Long Timelines
Evolution does not operate on a quarterly cycle, and neither does breakthrough innovation. Set horizons of 10 or more years for overall portfolio performance, make at least 25-30 investments before drawing conclusions about your strategy, and track metrics such as founder recycling and technology repurposing.
Case Studies in Evolutionary Pivoting
The Netflix Radiation
Netflix made some major pivots: from DVD rentals to streaming, from streaming to original content, and then from original content to gaming. Each tenfold revenue jump coincided with shifts in the external environment:
1. Broadband penetration exceeding 50% of households led to the rollout of streaming in 2007.
2. A 140% increase in content licensing costs between 2010 and 2013 drove the investment in original content in 2013.
3. Market saturation in 2022 prompted the creation of a gaming subsidiary.
These adaptations generated a 29,900% return from 2002 to 2022, outperforming 99.97% of public companies.
Synthetic Biology’s Industrial Evolution
Modern synthetic biology startups illustrate the industrial application of evolutionary principles. Companies like ALEph Bio expose yeast strains to simulated conditions in 10,000-liter bioreactors, evolving novel metabolic pathways that boost biofuel yield by 300% in just 11 weeks. They achieve in eight weeks what used to take 18 months, an 11.25-fold acceleration in generating variation. Similarly, Instacart pivoted from an “Amazon Fresh killer” to a white-label infrastructure provider during the COVID-19 crisis, showing how market pressures drive rapid adaptation.
The Resilient Innovation Ecosystem
The most dynamic startup ecosystems share key evolutionary traits:
• Low cost of experimentation: Silicon Valley makes it relatively affordable to start a company through shared infrastructure and services.
• Minimal stigma for failure: In the Bay Area, a failed venture is seen as a learning experience rather than a mark of shame, unlike regions where bankruptcy carries lasting stigma.
• Rapid resource recycling: In Silicon Valley, 68% of failed startup assets such as intellectual property, talent, and data are reused within six months compared to only 12% in Europe.
• Dense information networks: Knowledge about what works and what does not spreads quickly through meetups and accelerators, enabling information to diffuse up to five times faster than in more closed ecosystems.
These features create what biologists call optimal conditions for adaptive radiation, the rapid diversification of species to fill new niches. For startups, this means the ability to quickly spot and seize new market opportunities.
Building the Future Through Evolution, Not Design
In complex systems like markets and ecosystems, top-down design often falls short. The most powerful innovations emerge through evolutionary processes of variation, selection, and amplification. This perspective shifts how we approach innovation policy, investment strategy, and entrepreneurial education:
• For policymakers: Instead of picking winners, focus on reducing the costs of experimentation and failure. Consider measures such as bankruptcy laws that limit founder liability (similar to the UK’s 2024 Startup Act) to prevent ecosystem toxicity.
• For investors: Build portfolios that capture the value across a range of possibilities. The most successful VCs reject 99% of pitches yet maintain a 30% failure rate compared to 70% for “spray and pray” investors.
• For founders: See your venture not as a fixed entity but as an evolving organism that must adapt to survive.
• For ecosystem builders: Create environments that promote rapid learning cycles and efficient resource recycling. Innovation hubs like Station F and Kendall Square act as magnets for pioneering founders.
The Future of Evolutionary Entrepreneurship
Emerging tools now allow for more deliberate ecosystem engineering:
• AI-driven adaptation: Tools such as OpenAI’s FailureGPT can predict pivot success probabilities using over 130 adaptation signals.
• Cross-domain innovation: Initiatives like NEOM’s $500 million fund are designed to encourage collisions between biotech and fintech startups.
• Accelerated evolution: AI-driven A/B testing of over 100 million startup hypotheses annually is on the horizon for 2030.
At current mutation rates, with $468 billion in global VC investment in 2024, we might see the first “instant unicorn” via automated pivot systems reaching a $1 billion valuation in under 90 days by 2035, and biohybrid startups combining engineered organisms with AI optimization may emerge by 2040.
Conclusion: Survival of the Adaptive
The next revolutionary company is unlikely to emerge fully formed from a brilliant business plan. It will evolve from the remnants of earlier attempts that appeared to fail. The true innovation of the startup ecosystem lies in its capacity to turn failure into valuable information. As synthetic biology ventures demonstrate, directing evolutionary pressures through automated systems can compress innovation timelines from decades to quarters.
For all stakeholders, this means embracing three key principles:
1. Portfolio Speciation: Allocate over 20% to experimental areas outside your current focus.
2. Failure Metabolization: Build systems that quickly integrate components of failed startups, ideally within 90 days.
3. Cross-Kingdom Symbiosis: Encourage interactions between AI, biotech, and material science ventures.
So when someone raises concerns about the risk in venture capital, I explain that what appears as risk is really a sophisticated system for navigating uncertainty. Those apparent failures are not wasted capital; they are essential experiments that eventually lead to breakthroughs that change our world. Just as mitochondria merged with proto-eukaryotic cells to spark a revolution in life, the startups that will transform our century are already evolving from the remnants of today’s failed experiments. The lesson from 4 billion years of evolution is clear: extinction is the furnace in which innovation is forged. This is not a bug in the system; it is exactly how innovation works.
—
This is the piece I’m most proud of. When I was 14, biology was my favourite subject, but I was too busy trying to impress a girl whose name I don’t even remember, look cool to my schoolmates whom I no longer speak to & trying to figure out how to make money. I couldn’t see how my favourite subject would evolve into a favourable bank account, so I dropped it. Over the past four months, I’ve poured my heart into this piece, and it’s been a joy connecting with biology students from the University of Barcelona, Stellenbosch, and Oxford. Huge thanks to Vanina, Grainne, Rahul, Gregory, Vlad, Helenna, and Anya for putting up with my endless emails and video calls. I admire your courage in pursuing work that makes our world better. If you ever need anything, you have all my details.