Intel (INTC) Stock: Nvidia Taps Chipmaker for 2028 GPU Manufacturing

TLDR

• Intel stock gained in premarket trading after reports Nvidia will use Intel foundries for 2028 GPU production
• Nvidia plans to manufacture input/output die components using Intel’s 18A or 14A process nodes
• Intel will complete 25% of packaging work while TSMC handles 75% of Feynman GPU production
• The deal follows Nvidia’s $5 billion Intel stake purchase in December
• Partnership helps meet U.S. manufacturing rules while maintaining TSMC relationships

Intel shares moved higher Wednesday morning after DigiTimes reported Nvidia plans to use Intel’s manufacturing facilities for its next-generation chip.

Intel Corporation, INTC

The report said Nvidia will partner with Intel on the Feynman GPU architecture. This chip is set to launch in 2028.

Apple will also participate in the collaboration. The three companies are focusing on low-volume production for non-core components.

The arrangement allows the companies to comply with U.S. chipmaking regulations. It also preserves their working relationship with Taiwan Semiconductor Manufacturing Co.

Intel to Handle Specialized Components

DigiTimes cited supply chain sources saying Intel will manufacture portions of the Feynman GPU’s input/output die. This component manages communication between processor cores and external hardware.

Intel’s 18A or 14A manufacturing processes will be used for production. These represent Intel’s most advanced foundry technologies.

The company will also handle packaging for up to 25% of the chips. TSMC will manage the remaining 75% of packaging operations.

TSMC will still produce the majority of the Feynman GPU. The core silicon will use TSMC’s A16 process node, accounting for roughly 75% of the chip’s total value.

Investment Provides Boost for Struggling Chipmaker

The partnership news comes months after Nvidia invested up to $5 billion in Intel shares. That purchase happened in December 2024.

The investment joined financial support from SoftBank and the U.S. government. Intel has needed these capital infusions as the company faces mounting challenges.

Intel has lost market share to rivals in recent years. A series of strategic mistakes has damaged the company’s financial position.

The chipmaker released weak guidance last week. Intel said it cannot meet strong demand for AI server chips while its new PC processors hurt profitability.

Advanced Memory and Packaging Technology

The Feynman GPU will support next-generation memory standards. The chip may use HBM4e or HBM5 memory technology.

These memory types enable chips to process trillion-parameter AI models. Memory capacity per package will increase compared to current designs.

Intel will use its EMIB technology for the project. This system connects different chiplets within a single package.

Final assembly will occur at Intel facilities in the United States. This meets domestic manufacturing requirements for certain government and enterprise customers.

The Feynman architecture succeeds Nvidia’s current Rubin AI GPU lineup. Nvidia expects the new design to deliver performance improvements for AI workloads.

Intel’s participation gives the company a marquee customer for its foundry business. The chipmaker has been working to attract more external clients to its manufacturing operations.

Nvidia has not officially confirmed the partnership details. Intel also declined to comment on the DigiTimes report.

The post Intel (INTC) Stock: Nvidia Taps Chipmaker for 2028 GPU Manufacturing appeared first on Blockonomi.