AMD CEO Lisa Su said demand for the company’s Helios AI servers is strong as the chipmaker targets leadership in large-scale AI computing and inference workloads.
Advanced Micro Devices (AMD) has begun full production of its second-generation Helios artificial intelligence (AI) server platform, with shipments expected to start by the end of the third quarter. The company is positioning the platform as a key part of its effort to challenge Nvidia’s dominance in the rapidly expanding AI data centre market.
AMD chief executive officer (CEO) Lisa Su said customer demand for Helios is strong. The server platform combines AMD’s MI455X AI accelerators with its Venice central processing unit (CPU), both manufactured by Taiwan Semiconductor Manufacturing Company (TSMC).
OpenAI’s top computing executive, who joined Su during her keynote in San Francisco, said the ChatGPT maker plans to begin using Helios systems later this year.
AMD is targeting growing demand for AI inference, the computing process involved when AI models respond to user queries. Su said the company believes it can achieve leadership in scale-up computing, referring to systems designed to connect large numbers of processors and accelerators for demanding AI workloads.
AMD estimates that the global computing market could reach $2 trillion by 2030, including $1.4 trillion in AI acceleration chips and $220 billion in CPUs. The company estimated the market was worth $365 billion in 2025.
AMD is also expanding its AI partnerships. Cerebras Systems CEO Andrew Feldman announced a collaboration with AMD to combine their AI server technologies, initially for deployment in Cerebras data centres.
The chipmaker has separately announced plans to supply up to two gigawatts of its Instinct MI450 chips to Anthropic from the first half of 2027, alongside a potential investment of up to $5 billion in the AI company.
AMD is also working with cloud providers such as Vultr and TensorWave, which operate data centres using its hardware. The Helios launch highlights AMD’s broader strategy to compete with Nvidia across AI chips, server systems, cloud infrastructure and software.
