OpenAI and Broadcom Partner on AI Hardware - Latest News

OpenAI Secures New Hardware Partnership

OpenAI has established a new collaboration within the hardware sector.

The AI research organization revealed on Monday an agreement with semiconductor firm Broadcom for the provision of 10 gigawatts of bespoke AI accelerator hardware.

Deployment Timeline

These AI accelerator racks are scheduled for installation in OpenAI's own data centers, as well as those of its partners, beginning in 2026 and continuing through 2029.

According to a company statement, designing its own chips and systems allows OpenAI to integrate insights gained from developing advanced models and products directly into the hardware.

Financial Implications

Although the specific financial details of the agreement remain undisclosed, the Financial Times has projected a potential cost to OpenAI ranging from $350 billion to $500 billion.

Recent Infrastructure Investments

This partnership represents the most recent in a series of substantial infrastructure agreements for OpenAI.

Just last week, OpenAI announced a further purchase of six gigawatts of chips from AMD, a deal valued at tens of billions of dollars.

In September, Nvidia declared a $100 billion investment in OpenAI, accompanied by a letter of intent for OpenAI to utilize 10 gigawatts of Nvidia hardware.

Potential Oracle Deal

Reports also suggest that OpenAI entered into a significant $300 billion cloud infrastructure agreement with Oracle in September.

However, neither OpenAI nor Oracle has officially confirmed this particular arrangement.

TechCrunch has contacted OpenAI seeking further clarification.