Nvidia acquires SchedMD, raising AI fears over Slurm software bias

Published
Score
8

Why it matters

Nvidia acquired SchedMD, developer of the open-source Slurm workload manager, in December 2025 to bolster its AI and supercomputing ecosystem.[2][3] Slurm schedules computing tasks across hardware from Nvidia, AMD, Intel, and others, powering AI model training at companies like Meta, Mistral, and Anthropic, as well as government supercomputers for weather and security research.[2][3] Nvidia pledged to keep Slurm open-source and vendor-neutral, but AI executives, supercomputing specialists, and analysts worry it could prioritize Nvidia hardware in updates or roadmaps, eroding competition.[1][2][3]

Key players include Nvidia as acquirer, SchedMD as target, rivals AMD and Intel whose hardware Slurm supports, and users like Meta Platforms, Anthropic, and Mistral.[2][9] Analysts such as Manish Rawat of TechInsights and Dr. Danish Faruqui of Fab Economics highlight risks of "soft power" through delayed optimizations for non-Nvidia chips, citing Nvidia's 2022 Bright Computing acquisition as precedent where software favored Nvidia hardware.[2] Reuters reported concerns from five anonymous sources, including three AI industry insiders and two supercomputer experts.[2]

Originally from U.S. government labs, Slurm became essential for AI's compute-intensive workloads amid booming demand.[3] The deal closed late 2025, but scrutiny intensified with Reuters' April 6, 2026 report, amplifying fears of Nvidia dominating the software stack beyond chips.[1][2][8]

Newsworthy now due to AI's rapid growth and Nvidia's hardware dominance, the acquisition spotlights risks to open ecosystems and multi-vendor fairness, with experts watching future Slurm development for signs of bias.[2][3][6]

Sources

mail

Get notified about new Antitrust developments

Primary sources. No fluff. Straight to your inbox.

See more entries tagged Antitrust.