Trump’s AI Science Push Gets $100 Million Boost From a Computing Power Donation
The Trump administration is set to receive $100 million worth of computing credits from National Compute, giving its ambitious artificial intelligence science initiative access to additional high-powered computing resources as Washington seeks to accelerate research and compete with China in the field.
POLITICO reported that National Compute, a new company pooling computing capacity from major technology and infrastructure providers, plans to donate the credits to the administration’s Genesis Mission, according to people familiar with the plans.
The donation could help address the lack of access to computing power facing scientists hoping to use advanced AI. Training and running sophisticated AI models can require enormous amounts of processing capacity, while major cloud and infrastructure providers frequently reserve their most advanced systems for large technology companies and AI developers under long-term contracts.
The Genesis Mission, launched by President Donald Trump through an executive order in November 2025, is a government-wide effort to “train scientific foundation models and create AI agents to test new hypotheses, automate research workflows, and accelerate scientific breakthroughs.” The administration has set a goal of doubling the productivity and impact of U.S. science and engineering within a decade.
In July, the White House announced more than $5 billion in federal commitments to the Genesis Mission, expanding the initiative to more than 15 federal agencies. Projects include using AI to search for the causes of chronic diseases, accelerate drug development, analyze more than 150 petabytes of NASA data, design new materials, and develop autonomous laboratories.
National Compute’s contribution would add private computing resources to that effort.
The company is developing what it calls the National Compute Grid, designed to combine computing capacity that is currently scattered among cloud providers, research laboratories and other organizations. The system would include hardware using Nvidia and AMD processors as well as Google’s tensor processing units.
The concept resembles an electricity grid, but rather than every researcher or company building its own expensive computing infrastructure, users could tap available capacity across a larger network. National Compute aims to pool 2 gigawatts of computing capacity by 2030, with 750 megawatts allocated so far.
National Compute co-founder Anjney Midha said expanding access could allow far more AI companies and research groups to emerge. “We think there should be thousands more Anthropics and OpenAI’s,” Midha said, arguing that lowering barriers to computing capacity could allow scientists and researchers to concentrate on innovation instead of securing infrastructure.
The company is also rolling out a separate “Grid Defense” program intended to let federal agencies evaluate advanced AI models before deploying them. An AI agent called Marshall would coordinate testing while keeping government evaluation procedures confidential, reducing the risk that AI models could be trained specifically to circumvent those tests.
Government agencies would determine their own testing requirements, while researchers would receive anonymized results. National Compute says the system is designed to provide continuous monitoring of increasingly capable frontier AI models.
The program entered invite-only early access with some government agencies in June. National Compute now plans to allow users with .gov and .mil email addresses, along with students and faculty using .edu accounts, to request access.
Each eligible public-sector user will receive another $100 in compute credits, separate from the $100 million Genesis Mission contribution. The White House declined to comment on National Compute’s planned donation or its computing network.
