NVIDIA's Acquisition of SchedMD Raises Concerns in the AI Community, Potentially Leading to Performance Penalties for Other Chip Brands

According to 1M AI News monitoring, NVIDIA announced in December 2025 its acquisition of SchedMD, the primary developer of the open-source job scheduling software Slurm. Slurm is widely used for managing compute job scheduling in supercomputers and AI data centers, with approximately 60% of the world's supercomputers reportedly using it, including AI training clusters at companies like Anthropic, Meta, Mistral, and supercomputers used by various governments for weather forecasting and nuclear weapon development. Slurm is particularly adept at managing NVIDIA chips but has significant usage on non-NVIDIA hardware as well.

This acquisition has recently raised concerns among AI experts and supercomputer users who fear NVIDIA may subtly bias the software through updates towards its own chips (such as CUDA and InfiniBand), potentially impacting the performance of hardware from competitors like AMD and Intel. Some point to NVIDIA's previous acquisition of Bright Computing as a precedent, suggesting similar integrations could lead to performance penalties on other chips. Addison Snell, CEO of Intersect360 Research, stated concerns lie in NVIDIA potentially transforming this ubiquitous open-source tool into a product "more suited to or exclusive to its own hardware," affecting fair competition.

NVIDIA responded by stating that Slurm remains open-source software, and the company will continue to enhance and support it for all users, emphasizing its commitment to "open-source, vendor-neutral" development and stating that customers will broadly benefit. The company plans to maintain training and technical support for SchedMD's hundreds of customers while denying any past acquisitions have harmed multi-hardware compatibility. This event is seen as a critical test of whether Nvidia truly maintains openness in the AI and high-performance computing domains.