OpenAI's Infrastructure Expansion Continues

OpenAI anticipates further infrastructure agreements following significant deals with Oracle, Nvidia, and AMD. CEO Sam Altman is currently touring the Middle East and Asia to secure funding and partnerships to expand OpenAI's computing capacity. Altman met with major chipmakers, including TSMC, Foxconn, Samsung, SK Hynix, and Hitachi, to increase AI chip production and prioritise OpenAI's orders.

Samsung and SK Hynix are now memory-chip partners, projecting OpenAI's demand to reach 900,000 wafers per month, exceeding current global capacity. They plan to co-develop AI data centres in South Korea. OpenAI has partnered with Hitachi in Japan to build AI infrastructure. Altman is also seeking capital in the Middle East for projects like the Stargate data centre in Abu Dhabi, potentially supporting the deployment of Nvidia's Rubin systems in 2026. OpenAI expects to spend $16 billion on server rentals this year, potentially rising to $400 billion by 2029.

OpenAI is also shifting towards enterprise services, with collaborations focused on application integration, product development and infrastructure. SK Telecom and Samsung will build AI data centres in Korea, with Samsung SDS becoming a reseller for OpenAI's services. These moves align with the maturation of OpenAI's AI models and growing usage, with over 800 million weekly ChatGPT users and projected revenues of $13 billion in 2025.