On Thursday, Nvidia announced a $5 billion investment in Intel, supporting the struggling U.S. chip foundry but stopping short of granting Intel a vital manufacturing contract.
The agreement, which involves Intel and Nvidia collaborating on PC and data center chip development, could pose a threat to Taiwan’s TSMC. Currently, TSMC manufactures Nvidia’s top processors, business that might shift to Intel in the future. AMD, a competitor with Intel for data center chip supply, could also be impacted by Nvidia’s support.
Intel’s shares rose 12% in premarket trading, while Nvidia’s increased by 2%.
Nvidia, whose essential chips are driving a global AI surge, stated it will pay $23.28 per share for Intel’s common stock, slightly below Intel’s Wednesday closing price of $24.90.
This offer, however, is above the $20.47 per share that the U.S. government recently paid for a 10% stake in Intel.
Nvidia will become one of Intel’s largest shareholders, likely controlling 4% or more of the company after issuing new shares to finalize the deal.
Nvidia’s backing marks a new opportunity for Intel, following years of unsuccessful turnaround attempts by the renowned U.S. manufacturer.
Once the leader in the chip industry, Intel appointed Lip-Bu Tan as CEO in March. Tan is committed to making Intel’s operations efficient and expanding factory capacity only when demand justifies it.
Importantly, the deal does not involve Intel’s contract manufacturing, or “foundry,” business producing chips for Nvidia. Analysts believe Intel’s foundry needs a major client like Nvidia, Apple, Qualcomm, or Broadcom to survive.
This arrangement, however, adds to Intel’s growing capital reserves, following a $2 billion investment from Softbank and $5.7 billion from the U.S. government.
Intel CFO David Zinsner mentioned at a Deutsche Bank conference that the company was financially stable and wouldn’t need significant capital until there is strong demand for 14A, an upcoming manufacturing process.
HIGH-SPEED CONNECTIONS
Under Thursday’s agreement, Intel plans to design custom data center processors that Nvidia will pair with its AI chips, known as GPUs. A proprietary Nvidia technology will enable faster communication between Intel and Nvidia chips.
These high-speed connections are crucial in the AI sector, where numerous chips collaborate to process large data volumes.
Currently, Nvidia’s top-selling AI servers featuring these connections use only Nvidia chips, but the deal gives Intel an equal opportunity to profit from each Nvidia server.
The combined Nvidia-Intel chips could present a significant challenge to AMD, developing its own AI servers, and Broadcom, which offers chip-to-chip connection technology and aids companies like Google in developing AI chips.
In consumer markets, Nvidia will supply Intel with a custom graphics chip to pair with its PC processors, potentially giving it an edge over rivals like AMD.
While Intel’s x86 architecture has lost ground to Arm Ltd technology in data centers and PCs, it still holds the majority market share.
“This historic collaboration closely integrates Nvidia’s AI and accelerated computing stack with Intel’s CPUs and the extensive x86 ecosystem—a fusion of two leading platforms,” Nvidia CEO Jensen Huang stated. “Together, we will expand our ecosystems and lay the foundation for the next era of computing.”
Both companies did not disclose the financial terms of their technical collaboration but confirmed plans for “multiple generations” of future products. Nvidia and Intel described the partnership as a commercial arrangement to supply chips to each other, with no licensing involved.
No launch date for the first joint products was provided, but existing product plans remain unchanged.
In recent years, Nvidia has entered both the PC and data center processor markets. Meanwhile, Intel has been selling AI chips competing with Nvidia and aims to develop an AI data center server to rival Nvidia.
