Huang de Nvidia calls $ 5 billion stakes “an incredible investment”

Nvidia The CEO Jensen Huang said that the collaboration of investment and technology of $ 5 billion with the company with Intel comes after the two companies had discussions for almost a year.

Huang said he had communicated personally with the CEO of Intel Lip-Bu Tan about the partnership. He called Tan a “longtime friend” during a call on Thursday with journalists after companies announced that Nvidia co-develops the data center and PC chips with Intel within the framework of the investment agreement. On the call, Tan said that he and Huang have known each other for 30 years.

“We thought it would be an incredible investment,” said Huang.

Nvidia said he will collaborate with the flea manufacturer to create artificial intelligence systems for data centers that combine central processors based on Intel X86 with graphic processors and NVIDIA networking.

Intel Will also sell processors for PCs and laptops that incorporate NVIDIA, or GPU graphic processors.

The transaction itself took a few months to come together, wrote the Intel income manager Greg Ernst in a LinkedIn article, adding that the agreement had been concluded on Saturday.

The investment highlights the way in which the fortune of the two companies has commissioned the summit of the hierarchical order of the Silicon Valley following the explosion of the AI ​​inaugurated by the launch of Openai de Chatgpt at the end of 2022.

Intel shares are down 31.78% in the past five years, while NVIDIA shares have increased by 1,348% from opening prices on Thursday. Nvidia is worth more than 4.25 billions of dollars, while Intel is only worth $ 143 billion.

For decades, the most important part of a PC or a server was the central processor, and Intel dominated the market for these chips. But the infrastructure of AI, like the machines of the $ 4 billion data center, which Microsoft announced Thursday, often needs two or more Nvidia GPUs for each processor.

Nvidia AI systems, such as the NVL72 used by Microsoft, are delivered with Arm-CPU based, instead of processors based on Intel X86. During the call, Huang said that Nvidia will soon support Intel processors in his NVLink Racks for AI.

“We will buy these processors to Intel, then connect it to super chips which then becomes our calculation node, which is then integrated into an AI Supercalculator on Rack scale,” said Huang.

Nvidia will also contribute to GPU technology to Intel fleas that ship to laptops and PCs, which is a badly served market, said Huang. In total, the markets addressed for the two product collaborations are worth 50 billion dollars, said Huang.

“We are going to become a very large Intel processor client, and we are going to be a large supplier of GPU chips in intel chips,” he said.

Huang said the agreement with Intel will have “no” impact on Nvidia’s commercial relationship with Arm.

Thursday’s investment agreement focuses on the relationship between Nvidia and the Intel product division, not its foundry. The two companies, however, have not excluded future foundry partnerships.

“We have always evaluated the technology of Intel’s Foundry, and we will continue to do so, but today, this announcement is downright focused on these personalized processors,” said Huang. Nvidia currently uses Taiwan Semiconductor Manufacturing Company To make your chips.

Collaboration will use Intel packaging, which is a part of parts that occurs towards the end of the process and combines several chip components in one part which can be installed in the machines.

Tan said he was grateful to Nvidia’s vote of trust.

“” I would like to thank Jensen for self -confidence, and our team and Intel will work very hard to make sure it’s a good return for you, “said Tan.

Last year, the Intel board of directors deleted the previous CEO Pat Gelsinger due to the increase in costs in its manufacturing activity and the failure of the company to gain a foothold in ia chips. In March, Intel appointed Tan, a well -connected investor who had reversed the software company from Puce Cadené Design Systems, its new managing director.

Tan focused on reducing costs and fundraising in its short mandate, while the future of the company manufacturing the company, called Intel Foundry, remains vague.

In addition to the $ 5 billion in Nvidia and 8.9 billion dollars of the US government, Intel collected an investment of $ 2 billion in softbank, sold a majority participation in its subsidiary Asic Altera in Silver Lake for 3.3 billion dollars and sold $ 1 billion in Mobileye actions.

Intel has also reduced important staff, saying in July that they would eliminate 15% of its workforce by the end of the year.

The company develops its own chips and manufacture them. He wants to manufacture fleas for companies like Nvidia or AppleBut has not yet been secured as customers. Analysts say Intel needs a large foundry customer to point out that its technology is stable and ready for volume production.

But the manufacture of peak fleas is expensive and Intel reported that if it could not obtain enough customers, it may not continue to invest in its foundry. This could arouse a Washington reaction, whose politicians and lobbyists consider Intel as strategically important for the nation, because it is the only American company capable of manufacturing the most advanced fleas.

The Trump administration took a 10% stake in Intel in August. Intel was previously online to receive $ 8.9 billion in grants and loans from the flea law, but the Trump administration requested and received a participation in the flea manufacturer in exchange for money.

Huang was with Trump this week in England to attend a state dinner at the Windsor Palace and announce new projects and investments in the United Kingdom, but the Trump administration was not involved in this agreement, according to an official of the White House and Huang.

“Intel’s new partnership with NVIDIA is an important step for high-tech American manufacturing,” said White House spokesperson Kush Desai, in a press release.

– Megan Cassella de CNBC contributed to this story

WATCH: Nvidia wants Intel consumption activities, says Gene Munster from Deepwater