Anush Elangovan, corporate vice president of software development at AMD, says Brium’s advanced software capabilities will ‘strengthen our ability to deliver highly optimized AI solutions,’ including its Instinct GPUs that have been key to its competitive fight with Nvidia.
AMD is continuing its acquisition spree to challenge Nvidia’s AI computing dominance with the purchase of a startup that specializes in AI software optimization.
The Santa Clara, Calif.-based chip designer announced on Wednesday that it has acquired Brium, which is made up of “world-class compiler and AI software experts with deep expertise in machine learning, AI inference and performance optimization.”
[Related: Intel Spin-Off: Our InfiniBand Alternative For AI Data Centers Has A ‘Devastatingly Good’ Edge]
Financial terms of the deal were not disclosed.
Anush Elangovan, corporate vice president of software development at AMD, said Brium’s advanced software capabilities will “strengthen our ability to deliver highly optimized AI solutions,” including its Instinct GPUs that have been key to its fight with Nvidia.
“Their work in compiler technology, model execution frameworks and end-to-end AI inference optimization will play a key role in enhancing the efficiency and flexibility of our AI platform,” he wrote in a blog post.
The major advantage AMD sees with Brium is the startup’s “ability to optimize the entire inference stack before the model reaches the hardware,” according to Elangovan.
“This reduces dependence on specific hardware configurations and enables faster, more efficient out-of-the-box AI performance across a wide range of deployments,” he said.
Elangovan said the Brium team will “immediately contribute to key projects like OpenAI Triton, WAVE DSL and SHARK/IREE,” which are “essential to enabling faster, more efficient execution of AI models on AMD Instinct GPUs.”
“By focusing on new precision formats like MX FP4 and FP6, we are equipping our AI platform to handle emerging workloads in training and inference more effectively, helping developers achieve higher performance while maintaining efficiency and cost-effectiveness,” he wrote.
The Brium acquisition will also help “accelerate the open-source tools” underlying AMD’s AI software stack and advance its mission to support the specialized needs of customers in verticals ranging from health care and life sciences to finance and manufacturing.
“Their successful porting of the Deep Graph Library (DGL) to [the] AMD Instinct platform is a clear example of how they enable cutting-edge AI applications in health sciences,” Elangovan said. “This kind of domain-specific expertise enhances our ability to deliver optimized solutions for high-value industries, broadening our market reach and strengthening our position as a trusted partner across a diverse range of sectors.”
While AMD’s announcement and Brium’s website didn’t contain any details about the startup’s founding team, the LinkedIn profile for Brium co-founder and CTO Quentin Colombet said he has nearly a decade of experience of developing compilers and compiler optimizations for GPUs and accelerator chips at Google, Meta and Apple. Much of this work has been focused on machine learning.
Brium is the latest in a string of acquisitions AMD has made over the last couple years to boost its competitive stance against Nvidia, whose command of the AI market allowed the rival to earn more than double the revenue of AMD and Intel combined last year.
The most recent acquisitions include silicon photonics startup Enosemi and data center infrastructure provider ZT Systems, both of which closed this year. Prior to that, AMD acquired three software firms—Silo AI, Nod.ai and Mipsology—to improve its software stack.
“This is more than just a software play, it’s about delivering real value to customers, driving adoption of AMD platforms and helping define the next era of AI computing,” said Elangovan, who was founder and CEO of Nod.ai prior to joining AMD.
