Sentient Foundation Bets $42M to Keep AGI Open and Free

Sentient Foundation just committed $42 million to open-source AGI -- and the number is not an accident. The choice of $42 million is symbolic: the number is widely recognized as the fictional answer to "life, the universe, and everything," and for Sentient, it underscores the defining question of the AGI era -- who gets to build, access, and benefit from intelligence. The message is pointed, and the money is real.

The funding gap that nobody wanted to talk about

For the past several years, the AI industry has been moving toward an uncomfortable reality: the most powerful models, infrastructure, and capabilities are increasingly concentrated in the hands of a small number of well-funded companies. While billions of dollars continue to pour into proprietary AI platforms, the open-source ecosystem that helped drive much of the industry's early innovation has often struggled to access the same level of financial support.

While billions of dollars continue flowing into proprietary AI platforms, the open-source ecosystem has often lacked dedicated funding mechanisms capable of supporting developers, researchers, and startups at scale. Sentient Foundation is trying to change that with what it calls one of the largest commitments dedicated exclusively to supporting developers, researchers, and startups building artificial general intelligence technologies in the open.

Who is behind this

Sentient Foundation is a nonprofit dedicated to ensuring artificial general intelligence remains open, decentralized, and aligned with humanity's interests. The organization has deep academic roots: co-founder Pramod Viswanath is a professor at Princeton University, specializing in the first-principle understanding and design of blockchains and inventing communication algorithms via deep learning. The project previously raised serious capital: Sentient raised an $85 million seed round on a $1.2 billion valuation, co-led by Founders Fund (Peter Thiel's firm), Pantera Capital, and Framework Ventures.

The foundation has also shipped its own research. Sentient has contributed to open AI development through projects such as ROMA, Open Deep Search, EvoSkills, Arena, and its OML model family. OML -- which stands for Open, Monetizable, and Loyal -- is a cryptographic standard that is core to Sentient's thesis. OML allows developers to share access to their models (open) while being paid for that use (monetizable) and retaining control over how models are edited (loyal). Think of it as a way to make open-source AI economically sustainable without surrendering it to a platform.

How the money actually works

The program combines non-dilutive grants with founder-friendly investments, creating opportunities for teams at different stages of development. There are two distinct tracks:

• Grant Track: Intended for researchers, open-source maintainers, independent developers, and public-goods initiatives. Funding is provided without equity requirements, ownership claims, or restrictions on future development. Recipients retain ownership of their work and intellectual contributions.
• Investment Track: Intended for startups and teams developing commercial businesses around open-source AI technologies. Those investments will be founder-friendly, allowing teams to scale while retaining open-source access.

Beyond cash, the program also offers distribution support, compute credits, and hands-on engineering help from the team behind ROMA and Open Deep Search. Applicants will be screened based on technical merit, ecosystem impact, openness, and long-term potential.

What they actually want you to build

Sentient is not just writing checks to any open-source project. The grants page lays out a specific vision: AI that runs on hardware people actually own, that is private by default, and that is empowering rather than extractive. The foundation explicitly calls out use cases like medical guidance with no connectivity, scam protection for elderly users, multilingual tutoring tools, and crop-pricing tools for farmers. The framing is global and deliberately focused on markets that big labs ignore.

Rather than requiring applicants to fully open-source every component of their technology stack, Sentient's program takes a more pragmatic approach: teams must demonstrate that at least one critical element of their project remains openly available and contributes meaningfully to adoption, innovation, or ecosystem growth. This is a meaningful concession to commercial reality -- you can keep parts of your stack proprietary as long as the core contribution is open.

The timing is deliberate

For two years, open models trailed the closed ones by a wide margin. That gap has nearly closed. Open models are now within striking distance of the frontier. With recent releases like Sakana Fugu performing on par with Anthropic's Fable 5, the gap between open-source and closed AI development may now be only a matter of time. Sentient is betting that this is the window -- and that whoever funds the open ecosystem now shapes the defaults for the next decade.

Sachi Kamiya, Director of Venture and Growth at Sentient Foundation, put it bluntly: "The future of intelligence should be built by the many, not controlled by the few. A few companies are trying to become the OPEC of intelligence -- meter it, price it, decide who gets it. We're making it air."

Who is backing it

The initiative launches alongside support from a growing network of ecosystem participants, including Alibaba Cloud, Franklin Templeton, and leading academic institutions such as Princeton University and the Indian Institute of Science. Additional partners are expected to join as the initiative expands. An advisory council drawn from the open-source AI community will help guide funding priorities and review applications.

Who wins, who loses, and what changes

The clearest winners are independent researchers and small teams who have historically been priced out of the AGI race. Non-dilutive grants with no IP claims are genuinely rare in this space. Researchers, independent developers, open-source maintainers, and startups will receive support to build AI models, infrastructure, tools, and applications that remain accessible to the wider community.

The pressure lands squarely on closed-API providers. As foundation models become increasingly capable, a growing debate has emerged over who should control the future of intelligence. Many of today's most advanced AI systems operate behind closed APIs, with access controlled by a small number of providers. Critics argue that such concentration risks turning intelligence into a scarce resource controlled by a handful of corporations. A well-funded open ecosystem directly undermines that dynamic.

The second-order impact is geographic. The grants page explicitly calls out developers in Lagos, Bangalore, and São Paulo who can lose access to the models they depend on overnight. The Foundation argues that openness has consistently served as a catalyst for technological advancement -- the internet was built on open protocols, and much of the modern cloud computing stack runs on open-source software. Foundational technologies that power billions of users today often began as collaborative efforts maintained by global communities. The Foundation believes AI should follow a similar path.

How to apply

The application process has begun immediately under a continuous evaluation model to optimize capital distribution. There are no cohorts and no deadlines -- applications are open immediately and will be reviewed on a rolling basis. You can apply at sentient.foundation/grants. The foundation says it will reach deliberately into multilingual and underserved markets, and that projects with real repositories will move through review quickly.

The open-source AGI bet is no longer a fringe position. With closed models closing in on AGI-level capabilities and open alternatives narrowing the gap fast, the question of who funds the open ecosystem is no longer academic. "Open models are improving at an extraordinary pace," Kamiya said. "When they catch up, and they will, the people building on them should win, not pay rent on intelligence forever."