U.S. Senator Elizabeth Warren on Tuesday sent formal questions to the heads of the Department of Energy and the Department of Defense seeking details about government dependence on Nvidia following the chipmaker's acquisition of SchedMD, the company behind the Slurm workload manager.
Warren directed her letter to Energy Department Secretary Chris Wright and Defense Department Secretary Pete Hegseth, requesting information on how deeply each department relies on Nvidia's hardware and software offerings and whether either department has evaluated national security risks related to the takeover. The transaction that placed SchedMD under Nvidia's control closed in December, and the company did not disclose the financial terms of the deal.
Slurm is a scheduler used to allocate and manage computing tasks across clusters and supercomputers. The software is reported to be involved in operating about 60% of supercomputers worldwide. That role includes facilitating high-performance computing workloads for government systems that run forecasting, ballistic missile simulations, nuclear weapons development tasks and large-scale model training used in artificial intelligence projects.
Warren highlighted concerns that the acquisition could give Nvidia outsized influence over a critical software layer relied upon by rival firms and government users. In her letter she warned that the move "turns a once free software into one of NVIDIA's proprietary offerings, which may reduce competition and harm national security." The senator suggested Nvidia could gain control over a chokepoint that other firms need to run government supercomputers.
Some engineers and executives who depend on those supercomputer and AI systems have expressed unease that Nvidia might tilt the playing field in its favor, according to five people who have worked with such systems. At the same time, there are users who hope Nvidia's resources will accelerate long-awaited updates to Slurm, a codebase originally developed for government supercomputers.
"Customers everywhere benefit from our open source and free software," Nvidia said in a statement. "Slurm is open-source and we continue to provide enhancements for everyone."
Warren's letter also referenced Nvidia's prior purchases of other software providers that support computing infrastructure. She named Bright Computing and Run:ai as examples of companies that supply software critical to operating supercomputers and AI chip clusters. Nvidia acquired Bright Computing in 2022 and Run:ai in 2024; those deals, the letter states, have attracted antitrust scrutiny.
The senator argued that combining control of AI chips with critical software layers could enable Nvidia to make rival hardware harder to deploy and support, thereby disadvantaging competitors. "If NVIDIA controls both AI chips and the software layers that make these chips work, they are in a position to box out competitors by making their hardware harder to deploy and harder to support," Warren wrote.
Officials at the Energy and Defense departments were asked to disclose the extent of their dependency on Nvidia technologies and whether formal assessments of national security implications had been conducted following the SchedMD acquisition. The letter seeks responses and documentation that would clarify how government supercomputing operations might be affected.
The questions posed by Warren reflect mounting attention on how strategic software components and supporting services interact with hardware suppliers in an era where supercomputing and AI workloads are central to defense and research operations. The potential interplay between software control and hardware supply has emerged as an area of focus in discussions about competition and national security.
Key Points
- Senator Warren has asked the Energy and Defense departments to outline their reliance on Nvidia hardware and software following Nvidia's December acquisition of SchedMD, maker of Slurm.
- Slurm schedules computing tasks for roughly 60% of supercomputers worldwide and supports workloads used for defense-related simulations and large AI model training.
- The letter cites concerns that control of both AI chips and critical software layers could allow Nvidia to disadvantage competitors; Nvidia maintains Slurm is open-source and will continue to support it for all users.
Risks and Uncertainties
- Potential antitrust and competition concerns if control over Slurm results in reduced access or less competitive support for rival hardware - impacts technology and data center sectors.
- Uncertainty about whether the Defense and Energy departments have fully assessed national security implications of Nvidia's acquisition - impacts defense and government supercomputing operations.
- Possible operational risks for organizations that depend on Slurm if future software stewardship or support priorities shift - impacts research institutions, AI development groups, and cloud infrastructure providers.