Elon Musk says Tesla and SpaceXAI will build and run Terafab AI chip plant - Firstpost

Musk's push to move Tesla and SpaceXAI from chip design into fabrication is a structural strike on the foundry oligopoly's pricing power. Terafab, if it materializes, creates a third vertically integrated node in AI compute: upstream it locks ASML lithography slots, midstream it owns the wafer line, downstream it feeds Dojo clusters and FSD inference directly. For NVIDIA this is not competition—it is ecosystem erosion. When the end customer controls both design and manufacturing, the GPU premium thesis gets structurally undermined. On compliance, the CHIPS Act subsidy window aligns with the build cycle, but equipment sourcing from Taiwan, China or mainland suppliers triggers export-control review as a hidden cost multiplier. Over the next 18 months, expect NVIDIA to accelerate sovereign-customer LTAs to hedge substitution risk, Intel to weaponize its foundry arm by binding automakers, and TSMC's 3nm utilization to face its first genuine demand-side bargaining pressure. The long-tail: AI infrastructure converges into three closed loops, and the rent-a-chip model enters a slow deflationary channel.