NVIDIA Taps TSMC’s Next‑Gen Node to Keep Its AI Chip Lead

Recent reports indicate that NVIDIA is deepening its partnership with Taiwan’s foundry giant TSMC to manufacture its next generation of AI GPUs. The new chips are expected to target TSMC’s successor to today’s 3‑nanometer processes, aiming for significant gains in both performance and energy efficiency for AI data centers.

According to those reports, TSMC is carving out dedicated capacity for NVIDIA on its most advanced EUV production lines, while still trying to balance demand from other major customers. NVIDIA, for its part, is increasingly spreading manufacturing across several contract chipmakers, including TSMC and mid‑tier player UMC, as a hedge against supply disruptions and geopolitical risk.

With AI demand driving persistent GPU shortages and soaring electricity use, this tighter alignment between NVIDIA and TSMC could shape the global “AI infrastructure map” over the next few years. Observers also expect the new process technology to trickle down into lower‑power edge AI chips and system‑on‑chips for cars and robots, extending the impact well beyond hyperscale cloud data centers.