King of Hearts
The move would likely build on the Department of Energy's Genesis Mission, which has already pulled in over $800 million in private partnerships
The US government is set to announce an AI compute credit donation on Thursday. In a world where processing power has become the most fought-over resource in tech, that is a notable gift to hand out.
The announcement is widely expected to fit into the Department of Energy’s Genesis Mission, the federal effort to fuse national lab supercomputers with private cloud infrastructure. As of October 7, 2026, however, no official confirmation or details had been released ahead of the expected October 8 reveal.
What Genesis is and who has paid in #
The program was launched through an executive order in November 2025. Its stated goal is to speed up scientific work in energy, materials science, biology and national security.
By December 2025, 24 organizations had signed memorandums of understanding to participate.
By July 2026, partnerships tied to the mission had passed $800 million, including compute credits and infrastructure contributions.
Microsoft was one of the bigger names at the table. The company pledged $60 million in July 2026, with $40 million of that taking the form of Azure credits spread over three years for Genesis projects.
Google made its own move around the same time. It committed $40 million in AI tokens and cloud credits for DOE researchers.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
The idle hardware problem #
One day before the expected announcement, a separate coalition launched something called the National Compute Grid.
The pitch is simple. Lots of expensive AI hardware sits around doing nothing, and the coalition wants to pool that unused capacity across labs and cloud providers.
Utilization rates at many of these systems sit below 15%.
The grid is targeting an initial capacity of approximately 760 MW. The plan is to scale that to 2 GW by 2030.
Why compute became a policy issue #
Training and running advanced models requires enormous clusters of specialized chips. Researchers at universities and smaller labs often cannot get access, even when their science would benefit.
Genesis tries to close that gap by treating national lab resources and commercial cloud as one larger pool. Compute credits are the mechanism that lets a researcher tap that pool without owning the hardware.
What this means #
For researchers, a federal compute credit donation could matter more than a traditional grant. Money still has to be converted into hardware access, while credits skip straight to the part where experiments actually run. The size of Thursday’s donation, who gets access to the credits, and which providers supply the underlying capacity will all shape how meaningful the program becomes.
The overlap with the National Compute Grid is another thread. If public credits and private pooling efforts end up working together, the sub-15% utilization problem could start to shrink.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our