National Compute, a newly formed company pooling resources from major technology firms, plans to donate $100 million in computing credits to the Trump administration’s Genesis mission, a government-wide initiative to deploy artificial intelligence for scientific discovery. The company will announce the donation Thursday at an event with White House Office of Science and Technology Policy Director Michael Kratsios. The move marks a significant shift in how computing power, long concentrated among the largest AI labs, reaches smaller researchers and government scientists.
The donation highlights a growing divide in AI research access. Major technology providers typically reserve their most powerful computing resources for long-term contracts with the largest artificial intelligence companies, while smaller labs and research teams struggle to secure capacity for cutting-edge work. National Compute hopes to address this bottleneck by creating what it calls the National Compute Grid. Company documents describe it as functioning like “the power grid or interstate highway system”: a shared network connecting computing resources that currently sit apart from one another.
Who Gets Access, and Who Doesn’t
The National Compute Grid brings together capacity from multiple sources, including cloud providers deploying Nvidia and AMD processors and Google’s tensor processing units. Researchers, companies, and government agencies could purchase computing time as needed, reserve capacity for larger projects, or connect their own systems to the network. Users pay only for the power they actually consume. That’s meant to lower barriers for institutions without massive budgets.
Venture capitalists and institutional investors finance National Compute, including public investors such as universities and pension funds. Major chipmakers, cloud providers including Vultr and Crusoe, and universities back the company. By 2030, it aims to pool 2 gigawatts of computing capacity; 750 megawatts have already been allocated to date.
The $100 million donation specifically supports Genesis Mission researchers. Beyond that, National Compute says each public-sector user with a .gov or .mil email address, along with students and faculty with .edu addresses, will receive $100 in computing credits. A scheduling system opens access this week.
The Safety Questions Behind the Generosity
Alongside the computing donation, National Compute is launching “Grid Defense,” a program that lets government agencies test artificial intelligence models before deployment. Marshall, an agent embedded in the grid, coordinates testing while keeping evaluation procedures confidential. The company designed the system to prevent frontier AI models from learning how to evade safety checks.
Some government agencies entered invite-only early access in June. Grid Defense operates voluntarily: the government sees test results, but AI labs can choose whether to release a model. National Compute created the program in response to high-profile incidents involving AI agents, including attacks on companies that AI labs reported over the summer.
“This is continuous frontier alignment monitoring at scale,” said National Compute Cofounder Anjney Midha. The approach would “enable your public stakeholders to monitor you as they need,” he added. That framing emphasizes public oversight, though labs retain final authority over deployment decisions because participation is voluntary.
The Broader Context
The initiative aligns with the Trump administration’s stated approach to AI regulation, described as largely light-touch and focused on developing advanced models over robust guardrails, particularly in competition with China. The administration has called its research push a “golden age of science,” which Thursday’s event will celebrate. The White House declined to comment on the donation or the computing network itself.
Midha argued that expanding access isn’t merely a public service; it’s economically rational. “We think there should be thousands more Anthropics and OpenAI’s,” he said. “[B]ut to unleash that kind of innovation, which is extraordinarily profitable for the public, we need to reduce the barrier to entry, increase compute access, and allow the researchers and scientists to do what they do best.”
His argument casts computing access as a path to widespread prosperity. But it doesn’t address whether major technology and venture-backed companies’ concentrated ownership of physical infrastructure ensures genuine competitive entry, or simply lets smaller players rent access on terms set by the powerful.
Why This Matters:
Computing power has become essential infrastructure for scientific research and technological development. When the largest technology companies control it, universities, smaller research institutions, and public-sector scientists face barriers to entry. The National Compute Grid offers an attempt to democratize access—a principle worth supporting. Yet venture-backed private companies still serve as intermediaries, and Grid Defense relies on AI labs’ goodwill rather than binding requirements. As artificial intelligence increasingly shapes public health, safety, and national infrastructure decisions, who controls computing resources and how thoroughly those systems are tested before deployment remain questions with real consequences for democratic accountability and public safety.