The decision by the AI giant to stop moving forward with Intel's advanced manufacturing process signals deep challenges for the US push toward semiconductor independence.
In a significant development that reverberates through both Silicon Valley and Washington, Nvidia has reportedly halted its testing of Intel's advanced 18A chip manufacturing process. The move, confirmed by multiple sources in late December 2025, represents a stumbling block for Intel's strategy to reinvent itself as a global foundry for external clients and strikes a blow to broader US efforts to secure domestic supply chains for critical artificial intelligence technology.
The decision by Nvidia, currently the world's dominant designer of AI accelerators, to step back from the partnership evaluation caused immediate turbulence in the financial markets. According to reports from Bloomberg and Reuters, Nvidia had been testing the viability of manufacturing its chips using Intel's 18A process but has since "stopped moving forward." This development underscores the immense technical hurdles facing American manufacturers attempting to rival the established dominance of Taiwan Semiconductor Manufacturing Company (TSMC) in the era of generative AI.
Table of contents [Show]
The core of the issue appears to stem from technical performance metrics. While Intel has aggressively promoted its 18A node as a quantum leap in transistor density and efficiency, the reality on the production line has reportedly lagged behind expectations. According to TradingView, the tests have stalled specifically because "performance and yield targets fell short of expectations." In the semiconductor industry, "yield" refers to the percentage of functional chips on a silicon wafer; low yields make mass production economically unviable.
This technical bottleneck aligns with caution previously signaled by Intel's own leadership. As The Verge noted, Intel CFO David Zinsner admitted in October that 18A yields were "not where we need them to be to drive the appropriate level of margins." Crucially, Zinsner indicated that it could be 2026 or even 2027 before these metrics improve sufficiently. For a company like Nvidia, which operates on aggressive product release cycles to satisfy voracious global demand for AI infrastructure, waiting for yield stabilization appears to be a risk they are unwilling to take at this moment.
The immediate fallout was visible on Wall Street. Following the release of the reports, Intel shares slid significantly, dropping 2% in pre-market trading according to Investing.com. The market reaction reflects a crisis of confidence in Intel's "IDM 2.0" strategy-CEO Pat Gelsinger's bold plan to open Intel's factories to outside customers.
"The reported break raises new inquiries regarding the technical readiness of Intel's 18A process, which the company has been promoting as a key to its struggle against Taiwan Semiconductor Manufacturing Company." - TECHi Analysis
Financial analysts are now calculating the long-term cost of this setback. ValueSense reports that the suspension raises concerns about broader customer hesitation. If the industry leader, Nvidia, does not certify the process, smaller players may also pause their engagements. This hesitation risks the loss of expected revenue exceeding $1 billion and could hamper Intel's ability to scale its foundry business effectively.
Beyond the balance sheets, this development has profound implications for US technology policy. The Biden administration, through the CHIPS and Science Act, has poured billions into domestic manufacturing with the specific goal of reducing reliance on East Asian supply chains. Intel was positioned as the national champion in this effort.
Nvidia's decision to halt testing reinforces the current geopolitical reality: the most advanced AI chips will likely continue to be manufactured primarily in Taiwan for the near future. While Nvidia has invested in diverse partnerships, WinBuzzer reports that despite a $5 billion investment context, the technical realities of the 18A node have forced a pragmatic pause. This delay complicates the narrative of rapid US semiconductor independence, suggesting that the timeline for a fully strictly domestic supply chain for cutting-edge AI silicon is longer than policymakers had hoped.
Experts caution against viewing this as a permanent termination of the relationship. As discussed in hardware communities on Reddit, Intel is continuing to refine its roadmap, with future iterations like 18AP and 14A already in development. Furthermore, 24/7 Wall St. notes that Intel continues to highlight progress on the node for its own products, such as Panther Lake, which are ramping up toward production.
However, for the immediate future-specifically the 2025-2026 window-Nvidia appears to have signaled that it will not rely on Intel's foundries for its flagship products. The burden of proof now lies squarely with Intel to demonstrate that it can solve its yield issues and deliver a manufacturing process that is not just theoretically advanced, but economically viable for the world's most demanding tech companies.
An Indian tech CEO shares insights from 25 years of experience, arguing that AI is an amplifier, not a replacement. The article explores 10 real-world case studies where human empathy, creativity, and strategic thinking remain irreplaceable in business and society.
An Indian tech CEO argues that AI is an amplifier, not a replacement for humans. The article explores 10 real-world case studies, from medicine to leadership, where human empathy, creativity, and strategic thinking remain fundamentally superior to algorithms.
An in-depth look at why AI is a tool for human augmentation, not replacement. This article explores 10 case studies where human skills like empathy, strategic thinking, and ethical judgment remain superior, arguing for a future built on human-AI collaboration.
