OpenAI set to start mass production of its own AI chips with Broadcom, FT reports

OpenAI and Broadcom Form Strategic Alliance

OpenAI has announced a partnership with Broadcom to begin mass production of custom AI chips starting next year, marking a significant step in reducing its reliance on third-party suppliers and, most notably, NVIDIA[2][4]. This initiative is engineered to accelerate OpenAI’s progress in artificial intelligence and enhance the scalability and performance of its flagship tool, ChatGPT.

A Shift in the AI Hardware Landscape

• Broadcom’s Q3 2025 AI revenue surged to $5.2 billion, a 63% year-over-year jump, powered by its various AI accelerators and networking solutions[1].
• OpenAI expects its 2025 revenue to reach $12.7 billion as custom chip production ramps up, solidifying its disruptive market strategy alongside Broadcom[1].
• NVIDIA currently holds 88% of data-center AI chip revenues, but faces stiff competition as hyperscale clients like OpenAI seek custom solutions[1].

Technical and Market Implications

OpenAI’s move to produce its own chips is expected to shift the dynamics of the sector, as hyperscalers increasingly turn to custom silicon to meet specific workload demands. Broadcom, renowned for its expertise in designing analog components, networking, and mixed-signal subsystems, is steering the project with ambitious targets, including a $10 billion order commitment and shipments set to begin in early 2026[1][3].

This collaboration is fuelled by growing demand for AI accelerators and the forecasted $360 billion AI hardware market by 2032. The partnership aims to optimize inference workloads and cost-sensitive applications, gradually eroding NVIDIA’s dominance in the space[1].

Global Reach and Expansion

• Broadcom’s market spans semiconductors (58.4%) and infrastructures (41.6%), with sales distributed across the Americas (29%), Asia/Pacific (55.1%), and Europe/Middle East/Africa (15.9%)[3].
• OpenAI’s chip production is set for mass rollout globally, potentially reshaping how enterprises deploy AI solutions.

The Race to AI Hardware Innovation

As the AI hardware marketplace evolves, major players such as OpenAI, Broadcom, NVIDIA, AMD, Intel, and new contenders from China and beyond are accelerating innovation and redefining the competitive landscape. For enterprise clients and investors, this marks a pivotal moment to witness the next generation of AI accelerators and the ongoing battle for technological leadership[1].

As the AI hardware marketplace evolves, major players such as OpenAI, Broadcom, NVIDIA, AMD, Intel, and new contenders from China and beyond are accelerating innovation and redefining the competitive landscape. For enterprise clients and investors, this marks a pivotal moment to witness the next generation of AI accelerators and the ongoing battle for technological leadership[1]. As the AI hardware marketplace evolves, major players such as OpenAI, Broadcom, NVIDIA, AMD, Intel, and new contenders from China and beyond are accelerating innovation and redefining the competitive landscape. For enterprise clients and investors, this marks a pivotal moment to witness the next generation of AI accelerators and the ongoing battle for technological leadership[1].