Nvidia books TSMC 1.6nm process for Feynman in H2 2028 ⋆ Electronics Weekly
Electronics Weekly reports Nvidia has booked capacity at TSMC for its A16 1.6nm process for the Feynman post-Rubin GPU architecture. A16 is described as a successor to TSMC’s N2, with 8-10% more density and speed and 15-20% lower power. Production readiness is in H2 2026, with Feynman mass production in H2 2028, using SoIC and COUPE technologies.
How this was made
The 30-second read
Why it matters
If accurate, the booking provides incremental visibility into leading-edge node demand and Nvidia’s long-range GPU platform schedule, but the timeline is long and lacks financial quantification.
Market read
A forward-looking supply-chain and technology roadmap detail for Nvidia’s next GPU platform, with limited immediate trading impact due to the long ramp window.
What to watch
No disclosed contract size, yield assumptions, or customer volume ramp; execution risks in 1.6nm and SoIC/photonic integration could dilute the signal.
Background
The article frames Nvidia’s post-Rubin GPU architecture, Feynman, as moving to TSMC’s 1.6nm A16 process and advanced packaging/photonic technologies.
Ticker impact
Article says Nvidia booked TSMC A16 capacity for Feynman, with production readiness in H2 2026 and mass production in H2 2028.
Modestly positive for NVDA on supply-chain confidence, but likely limited near-term impact given the timeline to H2 2028.
The text is a forward-looking manufacturing allocation and technology roadmap detail, not a near-term revenue or earnings datapoint.
Market effects
Reinforces the AI GPU supply roadmap narrative around leading-edge nodes (1.6nm) and advanced packaging/photonic bandwidth.
Limited direct regional trading signal; mostly global semiconductor sentiment.
Highlights continued demand for leading-edge foundry capacity tied to AI accelerators.
Counterpoint
Capacity booking may be partially speculative or subject to change, so it may not translate into measurable near-term financial upside.
Key entities
- companyNvidia
Booked TSMC A16 capacity for its Feynman post-Rubin GPU architecture, with readiness in H2 2026 and mass production in H2 2028.
- companyTSMC
Provides the A16 1.6nm process capacity and related technologies (SoIC, COUPE) referenced in the article.
- product/architectureFeynman
Nvidia’s post-Rubin GPU architecture discussed as using A16 and advanced vertical stack and photonic bandwidth approaches.
