The auto industry is competing with AI for memory chips — and losing - Automotive News

As the artificial intelligence boom intensifies, the automotive industry is increasingly competing with tech giants for access to memory chips, a shift that is driving up costs and reshaping the suppl... Read original →

The AI compute arms race is starving automakers of access to advanced memory chips. With TSMC’s 3nm capacity—tightly coupled with EUV lithography—prioritized for NVIDIA, automotive-grade HBM and LPDDR5 face acute shortages. Technically, this forces OEMs to modularize software-defined vehicle architectures to reduce bandwidth dependency. Geopolitically, U.S. export controls and CHIPS Act subsidies fragment the supply chain; even Toyota’s Kentucky expansion can’t fully insulate it from cost inflation and lead-time risks. Mitsubishi’s repositioning will fail unless it anchors differentiation in bespoke EV platforms. NVIDIA, meanwhile, leverages its AI ecosystem to lock in wafer allocation, cementing de facto dominance in the automotive compute stack. Over the next 12–24 months, sustained high memory pricing will accelerate adoption of near-memory and compute-in-memory architectures—Tier 1 suppliers without dedicated foundry partnerships risk irrelevance in next-gen E/E systems.