Lisa Su, AMD’s CEO, announced a key shift in AI, signaling that the technology is now driving genuinely meaningful work. Her statement came shortly after AMD revealed a massive partnership with Anthropic, involving the deployment of up to 2 gigawatts of its latest Instinct MI450 GPUs. This deal, which includes a $5 billion equity investment, aims to power Anthropic’s AI safety models, with the first gigawatt expected online by mid-2027.
Su’s remarks mark a broader transformation in AI computing. For the first time in 2026, global AI inference compute the day-to-day use of AI models is set to outpace training compute, which focuses on building those models. Inference drives countless real-time applications, from ChatGPT queries to customer service bots, demanding vast, ongoing GPU power.
AMD’s strategic positioning in the AI race
This Anthropic partnership completes AMD’s strategic trifecta alongside collaborations with OpenAI and Meta, announced in October 2025 and February 2026 respectively. Together, these deals represent a significant chunk of AMD’s equity, reinforcing its role as a core infrastructure provider for the AI industry. Su also emphasized the increasing collaboration between hardware makers, AI developers, and cloud operators, pointing to a more interconnected ecosystem than ever before.
Investors should note how the rise in inference compute is reshaping demand for AI hardware. AMD’s growing GPU deployments align with a broader trend where cloud giants and AI firms ramp up spending on data centers and chips, echoing moves seen in the market like the AI data center spending surge and massive investments from rivals. This shift could redefine the competitive landscape, making AMD a critical player in powering real-world AI applications.
This article is for informational purposes only and does not constitute financial advice.
