Nvidia to Buy $5 Billion Stake in Intel, Giving Rival a Lifeline

The deal between the chipmakers, whose fortunes have diverged sharply, includes plans to collaborate on technology to power artificial intelligence.

Nvidia, the world’s dominant maker of artificial intelligence chips, said on Thursday that it would invest $5 billion in its struggling rival, Intel, in a deal that illustrates how booming demand for A.I. has reshaped the global technology industry.

The deal is a lifeline for Intel and highlights a reversal of fortune between the two Silicon Valley companies. Once an industry leader, Intel has struggled to keep up with technology shifts to mobile devices and A.I. Last month, the Trump administration agreed to buy a roughly 10 percent stake in the company to bolster its shaky financial position.

Nvidia, by contrast, has become geopolitically significant and the world’s most valuable public company, producing the key chips needed for developing A.I. Jensen Huang, the company’s chief executive, is an industry superstar and joined President Trump in Britain this week for a state dinner at Windsor Castle.

But as companies like Microsoft and Oracle pour hundreds of billions of dollars into A.I. data centers, Nvidia and Intel set aside their longstanding rivalry to make artificial intelligence chips that allow customers to tap into both companies’ engineering expertise and software systems.

News of the deal sent Intel’s shares soaring more than 25 percent to $31.79 on Thursday, while Nvidia’s stock rose about 3 percent. The value of the U.S. government’s stake in Intel, which it purchased for $20.47 a share late last month, also immediately increased.

The partnership between the two American chipmakers comes as China tries to reduce its dependence on U.S. chip technology and manufacture more semiconductors domestically. On Thursday, the Chinese technology giant Huawei said it was expanding A.I. chip development, which could challenge Nvidia.

As part of their deal, Nvidia and Intel said they would collaborate on developing chips for personal computers and data centers. Intel specializes in chips known as central processing units, that serve as the “brain” of a computer by coordinating different tasks across a machine. Nvidia makes graphics processing units, which are powerful for narrower tasks like crunching vast amounts of data.

The companies said that they would use Intel’s PC-era chip technology, known as the x86 architecture, which could appeal to companies that need huge amounts of computing power for data centers or research.

The deal will help Nvidia diversify amid rising competition from Broadcom, Amazon and others, said Handel Jones, a semiconductor consultant at International Business Strategies, which advises electronics companies.

In the past, Nvidia only sold custom chips with processors that relied on software developed by Arm, the British chip design company, Mr. Jones said. But many customers have built data center capabilities around Intel’s software and processors, known as CPUs. The agreement with Intel allows Nvidia to offer those businesses custom chips as well.

Mr. Huang said the partnership would allow Nvidia to enter “a market segment we’ve never addressed,” which he said could be worth $50 billion. “We’re going to build revolutionary products; first-of-its-kind products,” he added.

Mr. Huang said the agreement would not result in Nvidia manufacturing its chips at Intel’s factories. He also said the Trump administration wasn’t involved in the deal.

Nvidia, which has a market value of more than $4 trillion, said it would buy about $5 billion of Intel’s common stock, a stake worth about 4 percent of the smaller company. It will pay $23.28 per share, a discount from Intel’s closing price on Wednesday. Based on Thursday’s stock gains, Nvidia has already profited from the deal.

Lip-Bu Tan, the chief executive of Intel, called it a “game-changing partnership.” He said he has been focused on strengthening Intel’s financial position.

“The culture I want to have is really lean, fast-moving,” Mr. Tan said. He said Intel would benefit from working with Nvidia’s engineering team, which is known for moving quickly and developing the high-performing A.I. chips.

Nvidia and Intel have headquarters about two miles apart on an expressway in Santa Clara, Calif. But their paths diverged in recent years.

Intel, founded in 1968, was long one of Silicon Valley’s most powerful and consequential companies. It designed and manufactured the semiconductors that were essential for machines running Microsoft Windows that helped popularize personal computers. The company lived by the maxim of a long-serving chief executive, Andy Grove: Only the paranoid survive.

But over the past two decades, Intel lost its way, cycling through chief executives and corporate strategies, while laying off thousands of employees. It missed the smartphone revolution after Apple’s introduction of the iPhone in 2007. And it has struggled to capitalize on demand for A.I. applications.

Nvidia has prospered. Once a niche developer of chips used for video game consoles, the company’s chips are now being packed into data centers by the thousands to power chatbots and other A.I. applications for companies including OpenAI, Google and Microsoft.

The deal does not address Intel’s most pressing business challenges, including not having its own artificial intelligence chip, nor does it commit Nvidia to manufacturing chips at Intel factories.

Intel’s failure to attract manufacturing customers has forced it to delay the construction of a factory in Ohio, angering state and local politicians and raising questions about whether it should have received government funding to expand its manufacturing operations.

The Trump administration has stepped in to try to help Intel. In January, it approached Nvidia and asked it to invest in Intel, said two people with knowledge of the conversations, who were not authorized to speak about private discussions.

Early this year, Howard Lutnick, who had been nominated to become Commerce secretary, proposed spinning off Intel’s struggling manufacturing business, the people said. He planned to turn its operations over to Intel’s rival, Taiwan Semiconductor Manufacturing Company, and have Nvidia, Apple and other major chip buyers invest in the new company and have it make some of their semiconductors.

The talks fizzled after TSMC announced a plan to invest $100 billion in manufacturing over the next four years. But Intel’s troubles continued.

In August, the Trump administration’s deal for a roughly 10 percent stake in Intel’s business was made with money it had committed to give the company under the CHIPS and Science Act, a federal program signed into law in 2022 that delivered billions of dollars in grants to revive U.S. semiconductor manufacturing.

The investment included an incentive for the U.S. government to acquire 5 percent more of Intel at a discount if the company sells its majority ownership interest in its manufacturing business.

“Intel isn’t out of the woods yet,” said Patrick Moorhead, founder of Moor Insights & Strategy, a tech research firm. “They have a lot to prove.”

Related Content