As part of the new pact, Nvidia and Intel will jointly develop PC and data centre chips, though Nvidia stopped short of signing Intel’s contract chipmaking division to produce its GPUs. Instead, Intel’s foundry will supply central processors and advanced packaging for the joint products.
Nvidia CEO Jensen Huang clarified that the White House had not been directly involved in this partnership. “Microsoft and OpenAI have signed a non-binding memorandum of understanding (MOU) for the next phase of our partnership,” Huang told reporters, noting that the U.S. administration “would have been supportive.”
Industry analysts say winning Nvidia as a foundry customer remains critical if Intel is to revive its manufacturing arm. “This may be the first step of an acquisition or breakup of the company (Intel) among U.S. chip makers though it is entirely possible the company will remain a shadow of its former self but will survive,” said Nancy Tengler, CEO of Laffer Tengler Investments.
Nvidia will pay $23.28 per share for its stake, below Intel’s $24.90 closing price on Wednesday but still above the $20.47 price paid by the U.S. government. Nvidia shares rose 3.8% on the news.
Intel also secured $2 billion from SoftBank and $5.7 billion from the U.S. government, strengthening its cash reserves.
Experts warn the pact could put TSMC and AMD at risk. Taiwan’s TSMC currently manufactures Nvidia’s flagship processors, while AMD competes with Intel in supplying data centre chips. “AMD has been seizing market share in desktops and laptops for quite some time and this will help Nvidia out against its closest domestic peers, but I think TSMC may have the bigger risk to its operation over the long term,” said David Wagner, portfolio manager at Aptus Capital Advisors.
Intel’s new CEO Lip-Bu Tan, appointed in March, has promised leaner operations and disciplined factory expansion. The partnership with Nvidia is seen as critical to repositioning Intel as a serious player in the AI race.
The companies plan to design custom Intel processors that will connect seamlessly with Nvidia’s AI chips through a proprietary interconnect technology, enabling faster communication between CPUs and GPUs. Analysts say this could give Intel a stake in every Nvidia AI server sold, challenging Broadcom’s chip-to-chip solutions that currently dominate the market.