Key facts: Intel tests 18A multi‑die; backup AI foundry; Cadence 14A - TradingView

Intel is advancing its 18A process and advanced packaging technologies, exploring the feasibility of integrating four GPU dies into a single chip for NVIDIA's 2028 Feynman GPUs. This move aims to enha... Read original →

Intel’s push into 18A and multi-die integration is triggering a cascade across EDA, advanced packaging, and AI chip architecture. Its deep co-optimization with Cadence at 14A pressures rivals like Synopsys to accelerate process-design co-innovation. Geopolitical hedging drives NVIDIA and Google to treat Intel as a critical foundry alternative outside Taiwan, China—yet EUV access constraints could inflate yield ramp costs. TSMC will likely counter by locking in Apple and NVIDIA through exclusive CoWoS capacity and accelerating its Arizona fab expansion. If Intel delivers volume production by late 2026, it reshapes North America’s AI manufacturing base; failure relegates it to a politically favored but technologically secondary supplier.