TLDR
- Intel shares jumped in premarket trading after DigiTimes reported Nvidia plans to use Intel’s manufacturing for 2028 Feynman GPU components
- Nvidia will produce low-volume, non-core portions of the Feynman GPU’s input/output die using Intel’s 18A or 14A manufacturing processes
- Intel will handle up to 25% of final packaging for the Feynman chip while TSMC manages the remaining 75%
- The collaboration includes Apple and aims to comply with U.S. chipmaking regulations while maintaining relationships with TSMC
- Nvidia purchased up to $5 billion in Intel shares in December, providing crucial support as Intel struggles with strategic challenges
Intel shares climbed in premarket trading Wednesday after reports surfaced that Nvidia plans to tap the struggling chipmaker for its next-generation graphics processing unit.
Taiwan-based DigiTimes cited supply chain sources saying Nvidia will collaborate with Intel on portions of its Feynman microarchitecture. The chip is scheduled for release in 2028.
The partnership also includes Apple. The companies are targeting low-volume production runs for non-core components.
This strategy helps meet U.S. manufacturing requirements. It also keeps relationships with Taiwan Semiconductor Manufacturing Co. intact.
DigiTimes reported that portions of the Feynman GPU’s input/output die will use Intel’s 18A or upcoming 14A manufacturing technologies. The input/output die manages communication between processor cores and external components.
Intel will handle up to 25% of the final packaging requirements. TSMC will oversee the remaining 75% of packaging work.
Intel Gets Crucial Backing After Tough Stretch
The partnership comes as welcome news for Intel. The company recently received a $5 billion investment from Nvidia in December.
That investment joined support from SoftBank and the U.S. government. These backers have provided a lifeline as Intel faces financial pressure.
Intel has struggled with strategic missteps in recent years. The company has lost ground to competitors in key markets.
Last week Intel issued disappointing guidance for the current quarter. The company warned it can’t keep up with surging demand for AI server chips.
Intel’s new PC chips are also hurting profit margins. The company faces challenges on multiple fronts.
Manufacturing Plans Target U.S. Production
The Feynman GPU will still rely primarily on TSMC manufacturing. The core components will use TSMC’s A16 node, representing about 75% of the chip’s value.
Intel’s foundry will build the remaining silicon content. Final packaging will take place on U.S. soil at Intel facilities.
Intel will use its EMIB technology for chiplet interconnectivity. EMIB handles the bonding between different chip components.
The Feynman chip is expected to support HBM4e or HBM5 memory standards. These advanced memory types allow each chip to handle trillion-parameter AI models.
Nvidia currently offers the Rubin AI GPU. The Feynman architecture represents the next generation after Rubin.
Intel’s 18A and 14A nodes are among the company’s most advanced manufacturing processes. The partnership gives Intel a high-profile customer for these technologies.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
