Nvidia stock (NASDAQ: NVDA) tumbled nearly 1% on Wednesday after reports surfaced that the AI chipmaker has paused testing of Intel’s ambitious 18A manufacturing process.
The move marks a significant setback for Intel’s foundry revival strategy and reinforces questions about whether the chip giant can compete with Taiwan Semiconductor Manufacturing Company (TSMC).
Both Nvidia and Broadcom had been evaluating Intel’s 18A node over recent months, but the tests have stalled as performance and yield targets fell short of expectations.
What the 18A pause means for Nvidia’s supply chain
Intel’s 18A process promised significant improvements over its previous generation: 30% higher transistor density, 25% faster performance at the same voltage, and up to 36% lower power consumption for equivalent throughput.
These specs matter because every watt saved translates to lower data center cooling bills, and every percentage point of density improvement means more AI compute per chip die.
Broadcom and Nvidia spent months running qualification tests, essentially manufacturing small batches of experimental chips to see how the new process would perform in the real world.
The pause signals that those prototype runs didn’t deliver the goods.
When chip designers pull back from testing, it rarely means outright rejection.
Instead, it reflects either yield problems, performance shortfalls, or architectural mismatches that require more development time.
Broadcom had already flagged concerns, reportedly telling Intel engineers that 18A was “not currently viable for high-volume production” back in September.
Nvidia’s pause suggests similar issues remain unresolved.
For Nvidia, this is especially notable because the company operates on razor-thin timelines; new AI accelerators must launch on schedule or risk losing market share to competitors.
Sticking with proven TSMC capacity is the safer play.
Intel emphasized that 18A remains “on track,” but the reality is blunt: a pause from Nvidia, the world’s most influential chip customer, reads as a vote of no confidence.
Competing against a foundry that already has 72% market share and a 41% revenue surge in Q3 2025 requires flawless execution.
Intel has already pushed back the 18A timeline twice, first to 2026, then to mid-2026 due to IP qualification delays, and customer hesitation only compounds the delays.
Nvidia stock: Does this reshape the AI foundry map?
This development actually reinforces TSMC’s near-total dominance of advanced chip manufacturing.
The Taiwanese foundry controls nearly three-quarters of the market and is shipping the 3nm process at scale while maintaining 90% yields.
Samsung, Intel’s closest competitor, remains stuck with yield challenges of its own, even as it pushes its 2nm GAA (Gate-All-Around) technology.
For Nvidia and Broadcom, the path of least resistance is clear: keep working with TSMC.
The Nvidia partnership announced in September 2025, which included a $5 billion investment by Nvidia, was supposed to be that breakthrough moment.
But the deal carefully avoided any manufacturing commitments. Nvidia CEO Jensen Huang explicitly stated the company would continue relying on TSMC. The 18A pause confirms that reality.
For markets, uncertainty carries a heavier tax than bad news.
Intel shares fell 2-3% early Wednesday on the report. Nvidia, too, witnessed some correction, a sign that traders have already priced in the company’s dependence on TSMC.
Analysts like Angelo Zino from CFRA have warned that Intel’s foundry business will “continue to bleed cash at least through 2027,” while Samsung and TSMC will capture the lion’s share of AI chip production capacity.
The post Nvidia stock plunges after Intel’s 18A move: what does it mean for AI chips? appeared first on Invezz