Open-Source AI to Power US Plan to Double Scientific Output

SAN FRANCISCO, CA – April 27, 2026 – In a significant move bridging Silicon Valley innovation with national scientific ambition, the AI company Together AI announced today it has joined the U.S. Department of Energy's (DOE) Genesis Mission. The initiative is a historic national effort to leverage artificial intelligence to double the productivity and impact of American science and engineering within the next decade.

Together AI, a prominent advocate for open-source artificial intelligence, will contribute its high-performance computing platform and expertise to a project that brings together the nation's most powerful resources. The Genesis Mission unites the DOE's 17 National Laboratories with leading academic institutions and a formidable roster of private technology companies to build an integrated AI discovery platform. This platform will link the country's top supercomputers, experimental facilities, and vast scientific datasets to tackle critical challenges in energy, national security, and fundamental research.

A 'Manhattan Project' for the AI Era

The Genesis Mission, established by an Executive Order in late 2025, represents one of the most ambitious public-private technology partnerships in recent history. With a goal of establishing a new era of AI-powered discovery, the mission has been likened to a 'Manhattan Project' for its scale and urgency. It aims to create an "American Science and Security Platform" that serves as the central nervous system for this new research paradigm.

The initiative has an aggressive timeline, targeting an initial operating capability by August 2026. The DOE has already identified specific science and technology challenges across priority domains including advanced manufacturing, nuclear energy, quantum computing, and semiconductors. To fuel this effort, the department announced nearly $300 million in funding opportunities earlier this year to assemble interdisciplinary teams.

Together AI joins a veritable who's who of the technology industry in this endeavor. Other committed partners include giants like Google, Microsoft, OpenAI, NVIDIA, Amazon Web Services, and Anthropic, among more than 20 others. This broad coalition underscores the national strategic importance placed on maintaining leadership in both AI and scientific discovery, with the DOE stressing that any products developed for the mission will be "architecture-agnostic" to foster broad compatibility and prevent reliance on a single vendor.

The Open-Source Advantage in National Research

Within this star-studded consortium, Together AI's participation is notable for its staunch advocacy of open-source principles. The company argues that transparency and accessibility are not just beneficial but essential for accelerating scientific breakthroughs. This philosophy directly counters the trend of proprietary, closed-off "black box" AI models, where the inner workings are hidden from users.

"The Genesis Mission is exactly the kind of national effort that open-source AI was built for," said Vipul Ved Prakash, CEO of Together AI, in the company's announcement. "Science moves fastest when researchers can actually see inside the models they're working with, adapt them to their domain, and run them at scale. That's what we do, and we're proud to put that capability in the hands of the scientists tackling the hardest problems in energy and national security."

This approach is critical for scientific validity, allowing researchers to inspect, verify, and customize models for highly specific tasks—from simulating fusion reactions to discovering new materials. Together AI has been a significant contributor to the open-source community, releasing widely adopted techniques like FlashAttention, an algorithm that dramatically improves the efficiency and speed of large language models, and Mixture of Agents (MoA), a novel method that enhances accuracy by having multiple AI models collaborate to solve a problem. These technologies are directly applicable to the complex, large-scale computational challenges central to the Genesis Mission.

High-Performance Infrastructure for Frontier Science

Beyond its open-source ethos, Together AI brings its 'AI Native Cloud' to the partnership. The company operates one of the world's fastest AI inference platforms, an infrastructure purpose-built for the massive, time-sensitive workloads that frontier research demands. For scientists at the national labs, access to such a platform means the ability to run more experiments, analyze data faster, and iterate on hypotheses at a pace that was previously unattainable.

The company's deep roots in academia bolster its credibility in the scientific community. Co-founder Percy Liang directs Stanford University's Center for Research on Foundation Models (CRFM), a leading institution for open-source AI research. Meanwhile, researcher James Zou leads the AI for Science Lab at both Stanford and Together AI, pioneering breakthroughs in the application of AI to medicine. This research-driven culture is exemplified by projects like EinsteinArena, a platform where AI agents collaborate openly to solve complex scientific and mathematical problems, mirroring the collaborative nature of human scientific progress.

By providing a library of the latest open-weight models and the high-performance infrastructure to run them, the company aims to ensure that the national research community is not locked into proprietary systems as it builds the next generation of AI-powered science. This flexibility is key to fostering an ecosystem of rapid, decentralized innovation.

Navigating a New Era of Public-Private Collaboration

The Genesis Mission represents a new chapter in the relationship between the U.S. government and the private technology sector. By marshaling the innovative power of companies like Together AI alongside the immense resources of the national labs, the initiative aims to create a whole greater than the sum of its parts. This model, however, is not without its complexities.

Experts note that integrating open-source technology into sensitive national security and energy research requires a delicate balance. It necessitates robust security vetting and continuous monitoring to ensure the integrity of the systems, while also navigating complex questions around data sharing and intellectual property among public, private, and academic partners. The DOE's commitment to an "architecture-agnostic" platform is a crucial step in mitigating risks of vendor lock-in and promoting a more resilient and adaptable research infrastructure.

The inclusion of a dedicated open-source champion like Together AI is a strategic decision, signaling that the future of American science will be built not only on powerful technology, but also on the principles of transparency, collaboration, and shared progress. As the Genesis Mission moves forward, this partnership will be a critical test case for how open innovation can be harnessed to solve the nation's most formidable scientific and security challenges.