NVIDIA takes $5 billion stake in Intel, will collaborate on AI chips

NVIDIA,and Intel today announced a partnership to develop multiple generations of custom data center and PC products to accelerate applications and workloads in hyperscale, enterprise and consumer markets. NVIDIA to invest $23.28 per share in Intel common stockUS$5 billion (approximately RMB 35.5 billion).

This investment is subject to customary closing conditions, including necessary regulatory approvals.

That price is slightly lower than the $24.90 per share that Intel shares closed at on Wednesday and higher than the $20.47 per share that the U.S. government paid to acquire Intel's 10% stake last month.

NVIDIA will become one of Intel's largest shareholders, potentially owning 4% or more of Intel after issuing new shares to complete the deal.

This further increases Intel's capital reserves. In August of this year, SoftBank announced that it would invest in IntelUS$2 billion (about RMB 14.2 billion); immediately following the U.S. government's announcement that it would invest in Intel common stockUS$8.9 billion (about RMB 63.2 billion)The company owns 10% of Intel's shares, making it Intel's largest shareholder.

Intel's chief financial officer David Zinsner told investors at a Deutsche Bank conference last month that Intel's "cash position is good" and that it won't need much money until it sees huge market demand for its next-generation manufacturing process, Intel 14A.

Intel's U.S. stock rose more than 29% in pre-market, while NVIDIA shares rose more than 3%.

According to the announcement, the companies will focus on utilizing NVIDIA NVLink to seamlessly connect NVIDIA and Intel architectures - combining NVIDIA's AI and accelerated computing strengths with Intel's CPU technology and x86 ecosystem - to deliver top-notch solutions to customers.

In the datacenter, Intel will build NVIDIA's custom x86 CPUs, which NVIDIA integrates into its AI infrastructure platforms and makes available to the market.

In the personal computing (PC) space, Intel will build and bring to market x86 system-on-chips (SOCs) with integrated NVIDIA RTX GPU chipsets. These new x86 RTX SOCs will support a variety of PCs that require integrated world-class CPUs and GPUs.

"This historic collaboration tightly aligns NVIDIA's AI and accelerated computing stack with Intel's CPUs and vast x86 ecosystem, enabling the convergence of two world-class platforms. Together, we will expand each other's ecosystems and lay the foundation for the next era of computing." Jen-Hsun Huang, Founder and CEO of NVIDIA, talked about.

According to Jen-Hsun Huang, AI is driving a new industrial revolution and reshaping every layer of the computing stack - from chips to systems to software. At the center of this reinvention is NVIDIA's CUDA architecture.

Intel's x86 architecture has been the cornerstone of modern computing for decades, and Intel is continually innovating its portfolio to support the workloads of the future, Intel CEO Lip-Bu Tan talked about.

"Intel's leading data center and client computing platforms, combined with our process technology, manufacturing and advanced packaging capabilities, will complement NVIDIA's leadership in AI and accelerated computing to help the industry achieve new breakthroughs. We thank Jen-Hsun Huang and the NVIDIA team for their trust and investment in us and look forward to continuing to innovate and expand our business for our customers in the future." Lip-Bu Tan said.

Q&A

Q1: What is the exact amount and percentage of Intel's shares acquired by NVIDIA?

A: NVIDIA will acquire Intel shares for $5 billion at a price of $23.28 per share, and will hold about 4% of Intel shares after the acquisition is completed, making it one of Intel's largest shareholders.

Q2: What are the AI chip products that the two companies have collaborated on?

A: The collaboration includes two directions: in the data center space, Intel will build a new line of x86 CPUs specifically tailored to NVIDIA's AI infrastructure platforms; and in the consumer PC space, Intel will build x86 systems-on-chip integrating NVIDIA's RTX GPU chipsets, called "x86 RTX SoCs".

Q3: What role does NVLink technology play in this partnership?

A: NVLink is NVIDIA's interface technology for integrating the two companies' architectures to support data and control code transfers between CPUs and GPUs. Compared to other standards, NVLink enables faster inter-chip transfers, which is critical for AI applications that require multiple GPUs to work together.