OpenAI has been strategically structuring its infrastructure agreements, valued at over $1.5 trillion, with a small team of in-house experts, opting to sideline traditional advisors such as bankers and lawyers. This approach has allowed OpenAI to directly embed its expertise in developing advanced AI models into the hardware design, unlocking new levels of capability and intelligence.
These infrastructure deals include significant partnerships with companies like Nvidia, AMD and Broadcom. OpenAI is co-developing custom AI accelerators with Broadcom and deploying Ethernet-based networking systems. With AMD, OpenAI is planning to deploy up to 6 gigawatts of GPU compute capacity, starting with 1 GW in 2026. Nvidia is investing up to $100 billion in OpenAI to support the development of next-generation models. These partnerships aim to secure long-term demand for hardware, improve performance and efficiency, and foster innovation in AI data centre design.
OpenAI is also expanding its data hosting capabilities, including a major growth deal to expand into the UK. The company plans to establish 'Stargate UK' to deliver new AI infrastructure, deploying thousands of NVIDIA GPUs. This initiative will enable British businesses to store their data on UK soil, enhancing privacy, accountability, and national resilience against cyber threats.
