- OpenAI has entered a strategic partnership with Broadcom for the joint development of custom AI accelerators and integrated systems, with deployment scheduled to begin in late 2026.
- This collaboration is central to OpenAI’s massive infrastructure campaign, which includes major agreements with NVIDIA, AMD, and Oracle to secure an unprecedented total of 30.5 gigawatts of compute capacity.
- CEO Sam Altman is pursuing an ambitious, $10 trillion goal to build out 250 gigawatts of compute capacity over the next eight years.
(hightechPRIME.com) – OpenAI has taken a decisive step toward scaling its generative AI infrastructure with the announcement of a strategic partnership with global semiconductor leader Broadcom. This collaboration, which has been underway for the past 18 months, centers on the joint development of custom AI accelerators and integrated systems tailored specifically to OpenAI’s computational needs. Broadcom is scheduled to begin deploying OpenAI-designed hardware in the latter half of 2026, with a comprehensive rollout expected to be completed by the end of 2029. The agreement, reportedly valued in the multibillion-dollar range, will deliver 10 gigawatts of compute capacity—an unprecedented figure in the context of AI infrastructure.
This initiative forms part of OpenAI’s broader campaign to secure extraordinary levels of processing power in support of its expanding AI ecosystem. Under the leadership of CEO Sam Altman, the company has entered into a series of high-profile agreements with major industry players including NVIDIA, AMD, Oracle, Samsung, and SK Hynix. NVIDIA has committed $100 billion to supply 10 gigawatts of infrastructure, while AMD’s deal—estimated in the tens of billions—will contribute six gigawatts and may result in OpenAI acquiring a 10 percent equity stake in the chipmaker. Oracle’s Stargate Project will add a further 4.5 gigawatts to the mix.
Altman has reportedly articulated an ambitious long-term objective: to build out 250 gigawatts of compute capacity over the next eight years. This figure represents approximately one-fifth of the total energy generation capacity of the United States, which currently stands at around 1,200 gigawatts. Realizing such a goal would require an estimated $10 trillion in capital investment. To meet this challenge, OpenAI is exploring novel financing mechanisms, although specific strategies have yet to be disclosed. Despite its rapid user growth—now exceeding 800 million weekly users—the company’s financial trajectory remains uncertain. Projected revenue for 2025 is expected to reach $13 billion, a figure that falls significantly short of the funding required to support its infrastructure ambitions.
The scale and pace of OpenAI’s expansion have prompted scrutiny from analysts and industry observers, particularly regarding the strain such growth may place on global energy resources and the speculative nature of many AI-related investments. Nonetheless, OpenAI maintains that designing proprietary chips optimized for its models will yield substantial performance gains. Broadcom has echoed this sentiment, describing the partnership as a “pivotal moment” in the evolution of artificial intelligence.
Altman emphasized that the collaboration with Broadcom is essential to unlocking the full potential of AI technologies and delivering meaningful benefits to individuals and enterprises worldwide. By investing in purpose-built infrastructure and forging strategic alliances, OpenAI aims to position itself at the forefront of the generative AI revolution, even as it navigates the complex financial and operational challenges that accompany such transformative ambitions.
