In 2019, when Silicon Valley shrugged at language models, a former plant genomics researcher from Craig Venter's lab wrote a $50 million check to a small research nonprofit called OpenAI. His name is Samir Kaul. He has been right about transformative technology, repeatedly, for twenty years.
Before Samir Kaul was making calls that would reshape artificial intelligence, he was studying a weed. Not metaphorically. Arabidopsis thaliana - a small flowering plant of no particular agricultural significance - was the organism that occupied his early career at The Institute for Genomic Research, Craig Venter's pioneering lab outside Washington D.C. In December 2000, Kaul co-authored a landmark paper in Nature announcing the first complete genome sequence of a flowering plant. Volume 408, pages 796-815. Seven peer-reviewed papers deep into a scientific career.
That's the detail most profiles skip. They start with the investments. They miss the point. Samir Kaul's edge in venture capital is not pattern matching. It's the instinct of someone who has actually done the work - designed experiments, evaluated evidence, sat in a lab and tried to understand what something fundamentally is before deciding what it could become.
"I like technical risk and I don't take market risk."- Samir Kaul, Khosla Ventures
After earning his MBA from Harvard Business School, Kaul joined Flagship Ventures, where he co-founded Helicos BioSciences and served as the founding CEO of Codon Devices. These were not passive board seats. He built companies. He hired people. He made payroll decisions. Then, in 2006, Vinod Khosla was assembling the team for a new firm - one that would bet on technologies that felt impossible by the standards of the day. Kaul became a Founding General Partner.
The firm's thesis was simple and radical at once: back founders attacking problems too hard for most investors to evaluate. Clean energy. Advanced biology. Deep computing. The kind of work that requires you to actually read the technical papers, not just the pitch deck. Kaul was built for exactly this.
"You can only lose one time, your money, and that you can make infinite amounts."- Samir Kaul on asymmetric returns in venture capital
By 2019, Kaul had already backed companies that would become meaningful public enterprises - Guardant Health, pioneering liquid biopsies for cancer detection (NYSE: GH); Nutanix, reimagining enterprise cloud infrastructure (NASDAQ: NTNX); QuantumScape, the solid-state battery company (NYSE: QS). The thesis was working. But then came a call about a small San Francisco research lab that had been quietly building large language models. The lab was called OpenAI. Most investors were skeptical - language models were curiosities, not products.
Kaul wrote the first check. The largest check. $50 million when $50 million felt like a lot to spend on something you couldn't yet demonstrate to a skeptical LP base. The decision reflected everything about his method: this was technical risk - genuine uncertainty about whether the technology would work - in a market that was, if it worked at all, essentially unbounded. His kind of bet.
The rest is the history everyone is now scrambling to rewrite in their own favor. But Kaul was there first, and he was there because of how he thinks, not because of luck or timing.
"We look for great people trying to disrupt large markets with technology."- Samir Kaul on the Khosla investment framework
Today Kaul's portfolio reads like a catalog of problems that shouldn't be solvable yet. Varda Space Industries is building manufacturing facilities in orbit - the logic being that microgravity enables the production of materials impossible to make on Earth. Ultima Genomics is attacking the cost of DNA sequencing. Mainspring Energy is developing a linear generator that converts fuel directly to electricity. Impossible Foods is making meat from plants at scale. Mirvie is predicting pregnancy complications before they occur.
Each of these represents a genuine scientific or engineering frontier. That's not an accident. That's the methodology. When Kaul evaluates a company, he is evaluating whether the underlying science makes sense, whether the founders understand the actual mechanism of what they're building, and whether the market, if the technology works, is genuinely large. The market question is the easy one. The technical question is where most investors stop reading.
He doesn't.
Kaul's framework is structured around a distinction most investors blur: technical risk versus market risk. He will take the former. He avoids the latter. If a technology works, does a market exist? That's a question with a fairly clear answer, even for speculative technologies. Is the technology possible at all? That's the interesting question - and it's the one that requires a scientific mind to evaluate properly.
This is why his career path - from Craig Venter's lab to Harvard Business School to Flagship to Khosla - makes a certain kind of sense. He built the scientific fluency first, then learned the capital structure. Most VCs do it the other way around, or skip the science entirely.
The result is a portfolio that looks, in hindsight, almost obvious - but only because the technologies eventually worked. At the moment of investment, each one was genuinely uncertain. Liquid biopsies for cancer. Solid-state batteries. Plant-based meat at scale. Language models. Orbital manufacturing. The through-line isn't a sector. It's a method.
Kaul's non-consensus bets share a structure: high technical risk, large addressable market if solved, and a founder who understands the actual mechanism. He is particularly drawn to founders with deep scientific or engineering backgrounds - people who are not just disrupting a category but genuinely trying to solve a hard problem. The asymmetry he describes is real: you can lose your initial investment once, but if the technology works, the upside is functionally uncapped.
Samir and Puja Kaul both attended the University of Michigan - Samir graduating in 1995, Puja in 1998. In 2023, they donated $3 million to establish the Samir and Puja Kaul Directorship at the Michigan Institute for Computational Discovery and Engineering. MICDE sits at the intersection of computation and scientific discovery - exactly the space where Kaul built his career before moving to Sand Hill Road.
The gift wasn't random. Kaul has long believed that the bottleneck on scientific progress is not ideas but computation - the ability to model, simulate, and analyze at scale. Supporting MICDE is an investment in the infrastructure that produces the kind of founders he backs.
He chairs the Development Committee at UCSF Benioff Children's Hospital, one of the country's premier pediatric research institutions, and serves on the board of the Michael J. Fox Foundation for Parkinson's Research. These aren't trophy board seats. They are extensions of a genuine interest in how science translates into better outcomes for actual patients.
His first major scientific paper was about a weed. The Arabidopsis plant has no agricultural value. It has extraordinary value as a model organism for genetics research. Kaul understood the difference, and the logic shaped everything that followed.
He wrote a $50M check to OpenAI in 2019 - before ChatGPT, before GPT-4, before most of Silicon Valley stopped dismissing language models. The check was the largest in that round. He was also the first.
He roots for the Michigan Wolverines, Boston Red Sox, and Washington Commanders. Three franchises with very different relationships to winning. He has experience managing long-term conviction through volatile stretches.
One of his portfolio companies (Varda Space Industries) is building factories in orbit. The logic: microgravity enables production of materials physically impossible to make on Earth. Another bet where the market is obvious if the technology works.