AMD Aims to Compete With Nvidia’s AI Hardware Dominance By Acquiring Brium

AMD’s recent acquisition may weaken Nvidia’s dominance in the AI hardware space. On Wednesday, chipmaker AMD announced it had acquired Brium, a stealth-mode startup focused on AI software optimization. Financial terms were not disclosed.

According to a sparse blog post on Brium’s website, the company develops machine learning tools that enable AI inference—the process by which AI models interpret new data—across a range of hardware platforms.

In simpler terms, Brium’s technology helps adapt AI software to run on different types of hardware, potentially making it easier to move away from Nvidia-based systems.

A Step Toward an Open AI Ecosystem and Tackling Nvidia’s Dominance

AMD said the Brium acquisition supports its goal of building an open, high-performance AI software ecosystem. While AMD frames the acquisition as promoting an open AI ecosystem, it also targets Nvidia’s dominance in AI software design.

Brium’s only blog post, published in November 2024, highlighted the industry’s dependence on Nvidia and specifically mentioned AMD.

The blog notes recent progress in offering server-side inference alternatives to Nvidia hardware. AMD’s Instinct GPUs perform well, but developers struggle to fully leverage them due to Nvidia-optimized workloads. At Brium, our goal is to enable efficient model inference across various hardware architectures.

This is AMD’s fourth strategic acquisition in two years to support an open-source AI ecosystem, according to the press release. Prior acquisitions include Silo AI (July 2024), Nod.AI (October 2023), and Mipsology (August 2023).

Read the original article on: TechCrunch

Read more: U.S. Tightens Rules on Chip Design Software Sales to China