Nvidia is reportedly evaluating Intel’s cutting-edge 18A fabrication process, a move that could reshape the global semiconductor landscape. Sources indicate both Nvidia and Broadcom are running test chips on Intel’s 18A node, signaling growing interest in Intel Foundry Services (IFS) as an alternative to industry leader TSMC. If these tests prove successful, it could pave the way for manufacturing contracts worth hundreds of millions of dollars—and potentially mark Nvidia’s first meaningful move to onshore parts of its chip production.
Why This Matters
The semiconductor supply chain has been under intense pressure in recent years, with geopolitical tensions, trade restrictions, and supply chain disruptions forcing companies to rethink their strategies. Nvidia, heavily reliant on TSMC’s advanced nodes, particularly its N4 and upcoming N3 processes, has faced capacity constraints and increasing competition for TSMC’s top-tier manufacturing slots. By exploring Intel’s 18A process, Nvidia is diversifying its manufacturing partners, reducing reliance on a single foundry, and potentially insulating itself from geopolitical risks in Taiwan.
What Is Intel’s 18A Process?
Intel’s 18A (short for 18 angstrom, or 1.8nm equivalent) is one of the most advanced semiconductor fabrication processes in development. It’s the final node in Intel’s five-nodes-in-four-years roadmap and is expected to enter high-volume manufacturing in 2025. The 18A process introduces two key innovations:
- RibbonFET Transistors: Intel’s take on gate-all-around (GAA) transistors, which promise better control over current flow, reducing power leakage and improving performance.
- PowerVia: A backside power delivery network that separates the power and signal layers of a chip, potentially improving power efficiency and reducing resistance.
These innovations aim to make Intel’s 18A process competitive—or even superior—to TSMC’s upcoming N2 (2nm) node and Samsung Foundry’s SF2 process.
Nvidia’s Long-Term Strategy
Jensen Huang, Nvidia’s CEO, has previously acknowledged the possibility of working with Intel as a foundry partner. While Nvidia designs some of the most advanced GPUs and SoCs on the market, it relies almost exclusively on TSMC for fabrication. As demand surges for AI accelerators and data center GPUs like the H100 and upcoming B100, securing additional manufacturing capacity has become a top priority.
By testing Intel’s 18A process, Nvidia isn’t just kicking the tires. It’s taking an active role in determining whether Intel can deliver the performance and efficiency needed for future AI and data center workloads. This is particularly relevant as AI models grow exponentially larger and more complex, demanding ever-greater computational power and energy efficiency.
A Vote of Confidence in Intel Foundry Services
Intel CEO Pat Gelsinger has made turning Intel into a major foundry player a cornerstone of the company’s turnaround strategy. Winning over Nvidia—a company known for pushing the limits of chip design—would be a huge endorsement of Intel’s foundry capabilities. It would also signal to other fabless chipmakers that Intel is a viable alternative to TSMC and Samsung.
Broadcom’s reported testing of Intel 18A chips adds further validation to the process. Broadcom, like Nvidia, is a high-volume, high-performance chip designer that could benefit from Intel’s roadmap and U.S.-based manufacturing.
The Geopolitical Factor
The potential partnership also has geopolitical implications. With growing concerns over Taiwan’s security and the increasing pressure from U.S. lawmakers to onshore semiconductor production, Intel’s Arizona fabs could offer Nvidia a strategically important, U.S.-based supply chain option. The CHIPS and Science Act is funneling tens of billions into domestic semiconductor manufacturing, and Intel is among the largest beneficiaries.
What’s Next?
While Nvidia has denied involvement in discussions to buy a stake in Intel Foundry Services—as recently speculated—testing the 18A process is a concrete step toward a deeper relationship. If Intel’s 18A process meets expectations in these early tests, it’s likely Nvidia will seriously consider moving portions of its production to Intel fabs, especially for future AI chips where performance and power efficiency are paramount.
Key Takeaways
- Nvidia and Broadcom are testing Intel’s 18A process for potential future chip production partnerships.
- Intel’s years of development in 18A chip manufacturing technology could support high-end GPU production.
- The potential partnership could diversify Nvidia’s manufacturing options while bringing Intel significant new business.
Potential Partnership Dynamics
If Nvidia becomes a major customer for Intel’s 18A process, it could mark a new era in semiconductor manufacturing—and significantly shake up the balance of power in the global chip industry.
The possible collaboration between Nvidia and Intel using the 18A manufacturing process could reshape the semiconductor landscape. This partnership would leverage Intel’s advanced fabrication capabilities while helping Nvidia diversify its production options.
Strategic Implications for Nvidia and Intel
For Nvidia, testing chips on Intel’s 18A process represents a significant strategic shift. Currently, Nvidia relies heavily on TSMC for manufacturing its AI and graphics chips. Adding Intel as a production partner would reduce supply chain risks and create leverage in negotiations with TSMC.
Intel stands to gain tremendous credibility for its foundry business if it secures Nvidia as a customer. Pat Gelsinger, Intel’s CEO, has made expanding the foundry business a cornerstone of the company’s revival strategy. A deal with a high-profile chip designer like Nvidia would validate Intel’s manufacturing capabilities to other potential customers.
The partnership could also lead to technical collaboration that benefits both companies in the competitive AI chip market.
Intel’s 18A Process Node Innovations
Intel’s 18A process represents a significant leap in manufacturing technology. The process incorporates several key innovations:
- RibbonFET transistor architecture that improves power efficiency and performance
- PowerVia backside power delivery that reduces signal noise and improves efficiency
- Next-generation EUV lithography for creating extremely small circuit patterns
These technologies could help Nvidia build more powerful, energy-efficient AI processors. The 18A process is scheduled to be production-ready by late 2025, which aligns with Nvidia’s typical product development cycles.
Intel has invested billions in developing these advanced production techniques. The company claims 18A will be competitive with or superior to TSMC’s comparable nodes.
Industry and Market Impact
This potential partnership could reshape the global chip market. TSMC currently dominates high-end chip manufacturing, but Intel securing Nvidia’s business would signal a meaningful shift in the industry balance.
Stock markets have already reacted positively to rumors of the partnership. Intel shares jumped when news broke about Nvidia and Broadcom testing the 18A process.
For customers, more manufacturing competition could eventually lead to better chip availability and potentially lower prices. The semiconductor industry has faced severe supply constraints in recent years.
The deal might also influence other chip designers like AMD and Broadcom to consider Intel’s foundry services. Reports suggest TSMC has even pitched a joint venture with Intel to major customers, showing how this potential partnership is already affecting industry dynamics.
Production and Operational Considerations
As Nvidia explores using Intel’s 18A manufacturing process, several complex factors must be addressed before any formal production agreements can be established. The transition would involve significant planning for manufacturing capacity, rigorous testing protocols, and careful intellectual property arrangements.
Manufacturing Contracts and Capacity
Intel’s push into the foundry business faces substantial challenges in scaling up to meet potential demand from major chip designers like Nvidia. The company would need to guarantee sufficient wafer capacity to handle Nvidia’s production requirements without disrupting existing commitments.
Contract manufacturing agreements would likely involve multi-year commitments with specific volume guarantees. This creates financial risk for both parties – Intel must invest in capacity without certainty of continued orders, while Nvidia would become dependent on Intel’s manufacturing capabilities.
The timeline for production ramp-up would be a critical negotiation point. Intel’s Arizona facility has begun test production of 18A chips, but full-scale manufacturing requires additional investment and preparation.
Testing and Quality Assurance
The testing phase currently underway represents a crucial step in validating Intel’s 18A process for complex chips like those designed by Nvidia. These manufacturing tests evaluate multiple parameters:
- Process stability across different chip designs
- Yield rates for economically viable production
- Performance characteristics matching or exceeding current manufacturing partners
- Power efficiency meeting Nvidia’s specifications
Rigorous quality assurance protocols would need development before any production agreement. This includes establishing testing methodologies, acceptable defect rates, and remediation procedures for production issues.
For Nvidia, moving even non-critical products to Intel would require confidence in consistent quality across manufacturing runs.
Intellectual Property and Collaboration Framework
A framework protecting both companies’ intellectual property would be essential for any manufacturing partnership. Nvidia would need to share chip designs with Intel while maintaining ownership rights and preventing information leakage.
The potential collaboration faces additional complexity if TSMC’s reported joint venture materializes. Such an arrangement could create unique intellectual property sharing concerns among traditional competitors in the chip design ecosystem.
American manufacturing incentives might influence these arrangements, as government programs encourage domestic semiconductor production. These incentives could affect contract terms and manufacturing costs for complex chips.
Both companies would need clear boundaries around access to proprietary technologies while maintaining necessary information flow for successful production.
Frequently Asked Questions
Recent news suggests Nvidia and Broadcom are testing Intel’s 18A manufacturing process, potentially signaling a shift in chip fabrication partnerships. This development raises several important questions about the technical, strategic, and market implications.
What are the potential benefits for Nvidia in utilizing Intel 18A technology for chip fabrication?
Nvidia could gain access to cutting-edge manufacturing capabilities through Intel’s 18A process. This advanced technology might allow for better performance in high-end GPUs and specialized AI chips.
The 18A process could potentially help Nvidia create smaller, more power-efficient chips. This would be valuable for products that need to balance performance with energy consumption.
By working with Intel, Nvidia might also diversify its supply chain. This reduces risks associated with depending too heavily on a single manufacturing partner.
How could Intel’s 18A process impact the semiconductor market if adopted by industry leaders like Nvidia?
Intel could regain competitive position against TSMC and Samsung if major players adopt its 18A process. This would reshape the current manufacturing landscape dominated by these Asian companies.
The partnership might create new standards for chip design and production. When industry leaders collaborate, their practices often become benchmarks for others.
Consumer electronics might see improved performance and potentially lower costs if the 18A process proves more efficient than current options.
What strategic advantages might Intel offer to Nvidia over other fabrication partners?
Intel’s manufacturing facilities in the United States could appeal to Nvidia amid growing concerns about supply chain security. Geographic diversity in production can protect against regional disruptions.
Intel might offer favorable terms to attract Nvidia’s business as it works to grow its foundry services. This could include priority production slots or custom modifications to the process.
The companies’ complementary expertise could lead to innovations neither could achieve alone. Intel’s manufacturing knowledge combined with Nvidia’s chip design skills creates powerful possibilities.
In what ways could collaboration between Intel and Nvidia influence the future of GPU development?
Joint work might accelerate advances in GPU architecture that depend on manufacturing capabilities. Certain designs become possible only when fabrication technology reaches specific thresholds.
The partnership could lead to more specialized GPU designs for different markets. Working closely with a manufacturer allows for better optimization of chips for specific uses.
New cooling and power management solutions might emerge from the collaboration. Advanced manufacturing processes often require new approaches to these critical aspects of GPU design.
How does Intel’s 18A process compare technically to Nvidia’s existing fabrication processes?
Intel’s 18A process represents a significant advancement in manufacturing technology. Reports suggest it offers competitive performance compared to processes currently used for Nvidia products.
The 18A process features refinements developed over years of research at Intel. These include improvements to transistor design and manufacturing techniques that could benefit Nvidia’s chips.
Power efficiency might improve with the 18A process. This is especially important for high-performance computing applications where energy consumption is a major concern.
Could an Intel and Nvidia partnership in chip manufacturing present challenges to competitors like AMD or TSMC?
Such a partnership would create a formidable technical alliance in the semiconductor industry. Two leading companies combining strengths poses a clear competitive threat.
AMD might face increased pressure competing against Nvidia products made with Intel’s advanced process. This could force AMD to seek similar manufacturing partnerships.
TSMC might need to accelerate its own development roadmap to maintain its manufacturing edge. Competition from Intel would likely spur increased investment in advanced processes.
