Epic Move Shakes the Industry: NVIDIA Commits USD 5 Billion to Intel, Poised to Transform the AI Chip Landscape

As Intel transitions into a U.S. government‐backed enterprise, two leading American chipmakers have shifted from fierce rivals to strategic collaborators. On the evening of September 18, NVIDIA (NASDAQ: NVDA) announced a USD 5 billion investment in Intel (NASDAQ: INTC), acquiring common shares at USD 23.28 apiece—an infusion equivalent to roughly RMB 35.5 billion. Concurrently, the companies unveiled a partnership to co‐develop AI infrastructure and data‐center products that aim to accelerate deployment across hyperscale, enterprise, and consumer markets.

Jensen Huang described the agreement as “an incredible investment,” emphasizing that AI is driving a new industrial revolution and reshaping every layer of the computing stack, from silicon to systems to software. By integrating NVIDIA’s AI and accelerated‐computing platform with Intel’s CPUs and vast x86 ecosystem, the two firms will unite world-class platforms to establish the foundation for the next generation of computing.

This investment marks the first capital alliance between NVIDIA and Intel in Intel’s 57-year history, following Intel’s recent receipt of USD 5.7 billion from the U.S. government under the CHIPS and Science Act and a USD 2 billion strategic stake from SoftBank. Upon closing, NVIDIA could hold at least 4 percent of Intel’s shares, becoming one of its largest shareholders. That stake, alongside the U.S. government’s approximately 10 percent position, would be valued at about USD 14 billion—reflecting an unrealized gain of USD 4.9 billion since August’s initial agreement.

Analysts note that this equity infusion and deeper collaboration could offer Intel crucial support as it has struggled to defend its position in the high-performance chip market. While the deal does not extend to NVIDIA’s chip manufacturing, it strengthens Intel’s capital structure and opens significant avenues for joint development. Following the announcement, Intel shares leapt 27.6 percent at the open and closed up 22.8 percent—its largest single‐day gain since October 1987—while NVIDIA advanced 3.5 percent and TSMC rose 2.2 percent in U.S. trading.

By the September 17 close, NVIDIA’s market capitalization swelled by over USD 75 billion, and Intel’s by more than USD 27 billion, combining for a net increase of over USD 100 billion (roughly RMB 720 billion) in a single session. Bloomberg data shows Intel’s valuation has climbed to its highest level since the dot-com boom, with a forward price-to-earnings ratio of 57 times—the strongest since 2001.

Observers ask why these erstwhile competitors would forge such an alliance. Three factors emerge: both companies now occupy prominent roles in U.S. industrial policy; Intel’s x86 architecture complements NVIDIA’s AI-accelerator strategy; and the partnership lays groundwork for future mergers or business integration. Earlier in August, Intel disclosed that the U.S. government would assume a passive 10 percent equity position—financed by USD 5.7 billion in CHIPS Act subsidies and an additional USD 3.2 billion in federal support—making it Intel’s largest shareholder ahead of BlackRock. The government’s stake does not carry board representation or governance rights, though it will align with Intel’s board on major votes, heralding a shift from purely subsidizing to owning.

President Donald Trump hailed the agreement as a shrewd deal, noting that the government “got USD 11 billion of stock for free” and that allocating roughly 10 percent of Intel’s shares to the U.S. “was good for the company.” He indicated that such equity arrangements could recur under future industrial-policy initiatives.

NVIDIA’s departure from its earlier ARM-based Tegra CPU efforts underscores its renewed interest in securing custom CPU support. After its failed attempt to acquire Arm, NVIDIA recognized the need to supplement its GPU and DPU capabilities with a robust CPU partnership—particularly for consumer system-on-chip applications. Under the agreement, Intel will produce bespoke x86 CPUs for NVIDIA’s AI platform, while NVIDIA’s GPUs will integrate seamlessly via NVLink. For personal computing, Intel plans to deliver x86 system-on-chips embedding NVIDIA RTX GPU chiplets, enabling AI-capable PCs that merge world-class CPU and GPU performance.

Although precise financial details of the technical collaboration remain undisclosed, both companies have committed to co-develop “multiple generations” of products without exchanging licensing rights. Intel and NVIDIA will supply each other’s chips to power joint solutions, potentially posing a formidable challenge to AMD, which is also pursuing AI server architectures. Intel CEO Patrick Gelsinger affirmed that Intel’s decades-long x86 innovation, combined with its manufacturing and advanced packaging expertise, will complement NVIDIA’s leadership in AI compute and accelerated computing.

Yet NVIDIA’s focus on China remains pivotal. In April, U.S. export controls halted shipments of NVIDIA’s H20 and AMD’s MI308 AI chips to China. By mid-July, the White House pledged to approve H20 export licenses, and President Trump hinted at sharing 15 percent of China sales revenue with the U.S. government. While NVIDIA generated USD 44 billion in Q1 2025 revenue, China accounted for USD 5.5 billion—and in Q2, global revenue rose to USD 46.7 billion while China revenue fell 24.5 percent to USD 2.769 billion. Against this backdrop, NVIDIA projects China could deliver USD 50 billion in business this year, assuming roughly 50 percent annual growth, though domestic alternatives have emerged.

Reports indicate that Alibaba (09988.HK), ByteDance (unlisted), and Cambricon (688256.SH) have been asked to pause purchases of China-specific NVIDIA chips, including the yet-to-launch RTX Pro 6000D. Jensen Huang expressed disappointment but acknowledged broader geopolitical considerations. He underscored China’s role as a major AI research hub—nearly half of the world’s AI researchers are based there—and expressed optimism about obtaining export approvals for the forthcoming Blackwell (B30A) platform.

As U.S.‐China competition in AI compute intensifies, China’s leading technology firms are accelerating domestic chip development. At Huawei’s Full Connect Conference on September 18, the company unveiled its Ascend roadmap, with the next-generation Ascend 950PR set for Q1 2026 and subsequent chips targeting nearly doubling performance each year. Huawei also showcased its Atlas 950 and Atlas 960 supernodes, promising interconnect bandwidth and compute capacity that rival concurrent NVIDIA offerings.

Alibaba’s Pingtouge and Baidu’s Kunlun chips have already entered AI training scenarios, while Pingtouge’s latest AI PPU reportedly matches NVIDIA’s A800 and H20 performance. Zhang Wen, founder of Biren Technology, emphasized that China’s domestic large‐model advances drive demand for indigenous chips, accelerating a “de-NVIDIA-ization” trend across AI infrastructure. In summary, the unprecedented NVIDIA-Intel equity and technology alliance underscores a tectonic shift in the AI chip industry—melding two pillars of American semiconductor leadership even as geopolitical dynamics spur alternative domestic solutions.