Woofun AI reports that AMD has entered full production of its Helios rack-scale AI system, with shipments commencing shortly following announcements at the Advancing AI Conference. OpenAI’s infrastructure lead confirmed plans for large-scale Helios deployment and a partnership to develop the MI500 series AI chips alongside future products.

AMD is also collaborating with Cerebras to launch a high-speed inference solution later this year, combining Helios GPU server racks with Cerebras’ wafer-scale chips. CEO Lisa Su reiterated that the AI accelerator market could reach $500 billion by 2028, a scale comparable to the current total semiconductor market, with GPUs expected to capture the majority of share.