3-Minute Read | Private Research Labs: A “Third Way” for Science?
Feb. 19, 2026—A new research-and-development experiment is taking shape in San Francisco, backed by some of the most prominent names in technology. Episteme—funded by Sam Altman (OpenAI) and Masayoshi Son (SoftBank), among other investors—launched in late 2025 with a fully outfitted lab and an initial cohort of 15 scientists pursuing curiosity-driven work spanning artificial intelligence, biotechnology, energy, neuroscience, and materials.
The premise is straightforward: remove the time sink of grant proposals and publication pressure, and scientists will produce breakthroughs that eventually become enduring businesses. As CEO Louis Andre puts it, “Profit should be a byproduct… The main focus should be impact. When you do great science, commercialization opportunities will emerge.”
For UIDP members, Episteme is worth watching because it represents another explicit attempt to reimagine the research-to-impact pipeline—and raises questions about how industry, universities, and new intermediaries may collaborate or compete in the coming decade.
Why Episteme believes the current system is a bottleneck
Andre’s launch essay, Introducing Episteme: A New System for Science, argues that history repeatedly shows transformational ideas being overlooked, delayed, or under-supported. He highlights three bottlenecks: academia’s risk aversion, industry’s short-term incentives, and venture capital’s impatience for long-horizon science. Episteme positions itself as a “third path,” drawing inspiration from legacy corporate labs such as Bell Labs and Xerox PARC—bringing researchers in-house, funding them, and surrounding them with operational support from ideation through commercialization.
Implicit in this critique is a challenge to existing partnership models that rely on sequential handoffs—from basic research to sponsored projects to commercialization—rather than integrated, long-term collaboration.
People-first selection and built-in operationalization
Two details stand out in Episteme’s vision. First, Andre says researchers are selected heavily on “theory of change”—how they believe their work could reshape the world—rather than solely on past outputs. Failed ideas are treated as opportunities to pivot rather than as liabilities.
Second, Episteme emphasizes operationalization and business rigor as the missing link in translating science to application. The organization plans to provide centralized support for intellectual property, hiring, tax, and administrative infrastructure, allowing scientists to focus on research rather than grant writing or company formation. These functions mirror—and in some cases consolidate—capabilities that universities and companies often manage separately through technology transfer offices, sponsored research administration, and external partnerships.
The FRO Model: Mission-First Teams
A parallel “third-path” experiment is gaining traction in the U.S. and Europe: Focused Research Organizations (FROs). Unlike the researcher-centric vision of Episteme, FROs are designed to deliver specific, high-impact infrastructure—open tools, software, datasets, and experimental platforms—within a time-bounded window. Rather than a portfolio of loosely connected individual projects, the entire team collaborates on a single project. Operationally, FROs typically run as small, full-time teams (15–30 researchers) with budgets of roughly $30–50 million, usually over a five-year period. They borrow a startup mindset—clear deliverables, regular progress reviews, and shared goals.
The FRO model began in the U.S. with philanthropic support (including funders such as Eric Schmidt) for Convergent Research, which has launched eight FROs since 2021 exploring projects in mathematics, neuroscience, climate, biology, and artificial intelligence. The U.K.’s first FRO, Bind, launched with a blend of government and private funding to map drug-binding possibilities for disordered proteins associated with cancer and dementia—signaling growing diversification beyond philanthropy into public-sector participation.
The hard part: private money, long timelines
A Science.org analysis of Episteme’s model injects skepticism into the discussion. Author Adam Rogers notes that the recent history of billionaire-funded efforts to accelerate science is mixed. Fast Grants, a $50 million rapid-award program launched during the COVID-19 pandemic, produced publications but few clinical breakthroughs. Arena BioWorks, launched with $500 million and high-profile hires, reportedly folded after two years.
These cases highlight a recurring vulnerability: privately funded science initiatives must reconcile slow scientific timelines and shifting investor expectations. Rogers also challenges simplistic comparisons to corporate labs. While Bell Labs produced foundational discoveries such as the transistor and the laser, research by economist Ashish Arora suggests that many corporate labs ultimately emphasized applied development over basic discovery.
Why public science still matters
Studies cited in the Science.org article suggest publicly supported basic research continues to drive larger long-run productivity gains than private R&D alone. A 2024 study by Andrew Fieldhouse, an economist at Texas A&M University, estimated productivity returns from public, nondefense R&D funding at 140% to 210%.
In this context, Episteme’s model and FROs may function less as a replacement for public research systems and more as a complementary experiment, offering stability, infrastructure, and translation pathways in areas where promising work might otherwise stall. For universities and companies engaged in long-term collaboration, Episteme and FROs raise practical questions about where research, translation, and commercialization best occur, and how responsibilities may shift across institutions. As new intermediaries emerge, established partnerships may need to clarify their distinct value propositions, timelines, and roles within an increasingly diverse research ecosystem.
We want to hear from you. How might FROs and private research labs like Episteme complement—or complicate—existing university–industry partnerships? Share your thoughts on research funding models via our LinkedIn.
The 3-Minute Read is a UIDP member information piece and does not represent the opinions of our members or representatives. We welcome your comments on our LinkedIn profile.
Go Deeper
Andre, L. (2025, November 10). Introducing Episteme: A new system for science. Episteme (Substack). https://epistemescience.substack.com/p/introducing-episteme
Rogers, A. (2025, December 19). Silicon Valley bets on science—without strings. Science. https://www.science.org/content/article/silicon-valley-bets-science-without-strings
UIDP. (2025, November 4). Focused Research Organizations: A New Model for Building Scientific Infrastructure. https://uidp.org/3-minute-read-focused-research-organizations-a-new-model-for-building-scientific-infrastructure/
Vance, A. (2025, November 10). Exclusive: Altman and Masa back a 27-year-old’s plan to build a new Bell Labs Ultra. Core Memory (Substack). https://www.corememory.com/p/exclusive-altman-and-masa-back-episteme-louis-andre
