Intel's Foundry Strategy Under Scrutiny Amid Nvidia's 18A Testing Halt

The semiconductor industry is at a pivotal juncture, with Intel's foundry ambitions facing renewed scrutiny following Nvidia's decision to halt testing of the 18A process node. This development, coupled with broader financial and strategic challenges, raises critical questions about the long-term viability of Intel Foundry Services (IFS) in a market dominated by TSMC and Samsung.

Technical and Strategic Headwinds

Intel's 18A process, featuring RibbonFET gate-all-around (GAA) transistors and PowerVia backside power delivery, was heralded as a technical breakthrough capable of rivaling TSMC's 2nm node. According to , the 18A process promises a 25% performance improvement or 36% power reduction compared to Intel's 3nm node. However, TSMC's 2nm process, already in risk production by late 2024, offers a 10-15% performance boost and 30% lower power consumption over its 3nm predecessor. While Intel claims its 18A node will surpass TSMC's 2nm in performance-per-watt metrics (2.53 vs. TSMC's 2.27), yield rates remain a concern: Intel's 18A achieved 55% yield in Q3 2025, lagging behind TSMC's projected 60% for its N2 node.

The technical challenges are compounded by strategic setbacks. Nvidia's $5 billion investment in Intel in September 2025, while a symbolic win, did not include a commitment to use IFS for manufacturing. This decision has been interpreted as a vote of no confidence in Intel's ability to secure high-volume foundry contracts, a critical need for the company to justify its $20 billion capital expenditures on 18A production. Analysts at Forbes note that Intel's history of manufacturing delays-such as the delayed ramp of its 7nm node-casts doubt on its ability to maintain a competitive edge in the 2nm race.

Financial Realities and Customer Diversification

Intel's foundry business remains unprofitable, with IFS reporting a $2.3 billion operating loss in Q3 2025, despite a 54.8% operating margin improvement from -133% in 2024. Revenue from external customers remains negligible, with only $8 million of IFS' $4.2 billion Q3 revenue coming from third parties according to analysis. This lack of traction underscores the difficulty Intel faces in convincing clients like AMD and Apple to shift production from TSMC, which holds a 64.9% foundry market share in Q3 2025.

Customer diversification efforts, however, show incremental progress. Microsoft has committed to using the 18A process for its chip designs, and Tata Group and AMD are in discussions according to market reports. Intel's recent Foundry Direct Connect event, attended by 1,000 partners, highlights its push to build an ecosystem around its advanced packaging technologies like Foveros and EMIB according to industry analysis. Yet, these partnerships are not yet sufficient to offset the reputational damage from Nvidia's withdrawal.

Strategic Implications and Market Projections

The U.S. government's 9.9% equity stake in Intel, part of a $5.7 billion CHIPS Act investment, underscores the geopolitical imperative for domestic semiconductor leadership. However, this support does not guarantee commercial success. Intel's 18A node must attract major clients to validate its cost-effectiveness, particularly as PowerVia and RibbonFET technologies increase production complexity and expenses.

Market forecasts suggest the foundry sector will grow to $296.7 billion by 2030, driven by AI and 5G demand. For Intel to capture a meaningful share, it must secure long-term contracts with clients beyond Microsoft and Tata. The company's 14A node, a successor to 18A, is contingent on customer commitments, with CEO Pat Gelsinger emphasizing a "build-only-what-customers-need" approach. This cautious strategy, while prudent, risks delaying critical R&D investments in a rapidly evolving industry.

Conclusion: A High-Stakes Transformation

Intel's foundry strategy hinges on three pillars: technical differentiation, customer acquisition, and financial discipline. While the 18A process offers compelling performance advantages, its commercial success depends on Intel's ability to overcome yield challenges and attract high-volume clients. The recent setback with Nvidia highlights the fragility of its position in a market where TSMC's scale and Samsung's aggressive 2nm roadmap pose formidable barriers according to market analysis.

For investors, the key question is whether Intel can replicate its internal manufacturing turnaround in the foundry space. The company's $5 billion partnership with Nvidia, though not a foundry deal, demonstrates strategic flexibility in co-developing AI infrastructure. If Intel can leverage its PowerVia technology and government-backed capital to secure a 10-15% foundry market share by 2030, it could emerge as a credible alternative to TSMC. However, failure to attract major clients in the next 12-18 months may force a reassessment of its ambitious $20 billion capital plan.

In the end, Intel's foundry viability will be defined not by its technical prowess alone, but by its ability to execute on a business model that balances innovation with commercial pragmatism.