OpenAI is partnering with Broadcom to develop custom AI chips and systems for its infrastructure and partner data centers. The deal involves 10 gigawatts of chips and is valued at billions of dollars, with deployment starting in the second half of 2026 and aiming for completion by 2029. This follows previous agreements with NVIDIA and AMD for significant compute power. NVIDIA is investing $100 billion and providing 10 gigawatts, while AMD will provide six gigawatts, with the potential for OpenAI to acquire up to a 10% stake. OpenAI also secured 4.5 gigawatts of data center capacity from Oracle. CEO Sam Altman aims for 250 gigawatts of compute power in the next eight years, a massive increase. Achieving this would cost approximately $10 trillion, which would necessitate new financing strategies. Even with current deals, OpenAI faces immense financial commitments. Despite significant investments from companies like NVIDIA and Microsoft, no single entity can fully fund OpenAI's ambitions. OpenAI's projected revenue for this year is $13 billion, which is insufficient to offset the massive required investment. OpenAI's aggressive expansion plans highlight its significant need for compute power.