OpenAI Secures Multibillion-Dollar Chip Agreements, Excluding Key Industry Players

OpenAI has finalized a series of multibillion-dollar agreements with leading semiconductor manufacturers, including Nvidia, AMD, Broadcom, and Cerebras, to secure next-generation AI accelerators. The deals, valued collectively in excess of $12 billion over a three-year period, reflect the generative AI leader’s aggressive scaling of its compute infrastructure to support large language model training and deployment. The scale of the commitments highlights the growing centrality of specialized AI chips in the technology ecosystem. Nvidia’s dominance in the AI accelerator market is reinforced through a $7.2 billion contract, which includes exclusive access to upcoming H200 and B100 chip variants. AMD received a $2.8 billion agreement for its MI300X series, while Broadcom secured $1.5 billion for custom silicon and system-on-chip solutions. Cerebras, known for its wafer-scale engines, was awarded $1.5 billion for integration into OpenAI’s data centers. Notably, several prominent semiconductor firms were not included in the initial deals, including Intel and MediaTek, raising questions about their future positioning in the AI hardware supply chain. Analysts suggest the exclusions may signal a strategic pivot by OpenAI toward partners with deeper AI-optimized silicon and faster time-to-market capabilities. The announcement immediately impacted market sentiment, with NVDA, AMD, AVGO, and CRBC shares rising 4.5%, 6.3%, 3.1%, and 8.7% respectively in after-hours trading. The moves underscore the increasing concentration of high-performance chip demand in a select group of suppliers, potentially reshaping competitive dynamics across the semiconductor sector.