Intel's turnaround strategy suffered a major setback this week as reports emerged that Nvidia has declined to use Intel's 18A manufacturing process for its chips. The decision, first reported by Reuters, comes after Nvidia tested Intel's next-generation fabrication technology and decided against moving forward—a blow to Intel's aspirations to become a foundry of choice for leading chip designers.

The development is particularly stinging given that Nvidia invested $5 billion in Intel shares just last month, part of a deal to co-develop chips for data center servers and personal computers. That investment had fueled optimism about a deeper manufacturing relationship. Instead, Nvidia's foundry decision suggests the collaboration will be limited in scope.

What Is Intel's 18A Process?

Intel's 18A is the company's most advanced chip manufacturing technology, roughly equivalent to what the industry calls "2 nanometer" or "sub-2nm" production. The process represents Intel's attempt to leapfrog competitors and reclaim leadership in semiconductor manufacturing after years of falling behind Taiwan Semiconductor Manufacturing Company (TSMC).

Key features of 18A include:

  • RibbonFET technology: Intel's version of gate-all-around transistors, which improve power efficiency and performance
  • PowerVia backside power delivery: A technique that improves chip density by moving power delivery to the back of the chip
  • Aggressive density improvements: More transistors per square millimeter than Intel's current 20A process

Intel has staked much of its turnaround on 18A, investing billions in new manufacturing facilities and positioning the technology as the cornerstone of its IDM 2.0 strategy—the plan to manufacture chips both for itself and for external customers.

Why Nvidia Said No

The specific reasons for Nvidia's decision aren't entirely clear, but several factors likely contributed:

Yield Concerns

Chip manufacturing yield—the percentage of functional chips produced from each wafer—is critical for economics. New processes typically start with lower yields that improve over time. Nvidia may have concluded that 18A's current yields aren't sufficient for high-volume production.

Risk Assessment

Nvidia generates enormous revenue from its AI chips, primarily manufactured by TSMC. Switching any production to Intel represents significant risk. Even a successful transition would require substantial engineering effort and could disrupt supply chains.

Competitive Dynamics

While Nvidia invested in Intel, the companies remain competitors in certain markets. Nvidia may be cautious about deepening dependencies on a company that also designs its own products.

"Much of Intel's turnaround is riding on the success of its foundry business. It has yet to secure a major customer for the new 18A process."

— Industry analysis

The Foundry Challenge

Intel's foundry strategy faces several structural challenges that Nvidia's decision highlights:

TSMC's Dominance

TSMC manufactures the vast majority of the world's most advanced chips. Apple, Nvidia, AMD, and Qualcomm all rely on TSMC for their most critical products. This concentration creates network effects—TSMC gets more experience optimizing processes, which attracts more customers, which generates more experience.

Trust Deficit

Intel spent decades as a vertically integrated company, designing and manufacturing chips only for itself. Potential foundry customers may question whether Intel can truly prioritize their needs alongside its own product divisions.

Execution Questions

Intel's manufacturing fell behind TSMC partly due to execution problems with previous process transitions. While 18A represents a fresh start, customers remember the struggles with 10nm and 7nm processes.

Intel's Response

Intel has emphasized that its foundry strategy doesn't depend on any single customer. The company has highlighted:

  • Government support: Billions in CHIPS Act funding to support U.S. manufacturing
  • Upcoming products: Panther Lake chips for PCs and Clearwater Forest for data centers will showcase 18A capabilities
  • Other potential customers: Microsoft, Amazon, and other tech giants have explored Intel manufacturing
  • Apple rumors: Speculation persists that Apple could use Intel's 18A-P process for some lower-end chips

CEO Pat Gelsinger has acknowledged that Intel won't pursue "major buildout" of new facilities without customer commitments, suggesting a more cautious approach to foundry expansion.

Market Implications

For investors, Nvidia's decision has several implications:

Intel's Valuation

Intel shares rose 86% in 2025, partly on foundry optimism. The Nvidia setback could temper enthusiasm, though much depends on whether Intel can secure other major customers.

TSMC's Position

TSMC's dominance appears more durable than some hoped. The company's N2 (2nm) process is on track for 2025 volume production, maintaining its technological lead.

Nvidia's Supply Chain

Nvidia remains dependent on TSMC for its most important products. Any TSMC capacity constraints or geopolitical disruptions affecting Taiwan could impact Nvidia supply.

The Bigger Picture

Intel's foundry challenges reflect broader questions about semiconductor manufacturing:

Geographic Concentration

The world's most advanced chips are overwhelmingly made in Taiwan, creating geopolitical vulnerabilities. Intel's foundry push was partly intended to provide a U.S.-based alternative.

Capital Intensity

Leading-edge chip manufacturing requires enormous capital investment—tens of billions of dollars for new fabs. Only a handful of companies can afford to compete, limiting options for chip designers.

Technology Cycles

Each new process generation requires years of development and optimization. Intel's 18A represents a multi-year bet that won't pay off immediately even if execution is perfect.

What's Next for Intel

Intel's foundry strategy isn't dead, but it needs concrete wins to maintain credibility:

  • Near-term: Successful launch of Panther Lake and Clearwater Forest chips using 18A
  • Medium-term: Announcement of at least one major external customer for 18A production
  • Long-term: Demonstration that Intel can compete with TSMC on yield, performance, and economics

The $5 billion Nvidia investment provides capital but not customers. Intel still needs to prove that its manufacturing can meet the demanding requirements of the world's leading chip designers.

The Bottom Line

Nvidia's decision to pass on Intel's 18A manufacturing is a significant setback for Intel's foundry ambitions, but not necessarily fatal. The semiconductor industry moves slowly, and Intel still has opportunities to prove 18A's capabilities with its own products before external customers commit.

For investors, the message is that Intel's turnaround remains a work in progress. The stock's strong 2025 performance priced in considerable optimism about the foundry strategy. That optimism now needs validation through actual customer wins—something that remains elusive heading into 2026.