TSMC Earnings and Intel Foundry (www.chipstrat.com)

TSMC reported record revenues and told investors that AI-driven demand is "real and accelerating," prompting heavy investment across Taiwan, new U.S. fabs, and advanced packaging capabilities. CEO C.C. Wei emphasized TSMC’s unmatched visibility into customer roadmaps, allowing the company to scale capacity and prioritize the most in-demand process nodes and packaging technologies. For AI/ML this means more supply of cutting-edge wafers and multi-die packages optimized for large models — higher transistor density, tighter power/performance envelopes, and broader availability of high-bandwidth memory integrations and chiplet-based designs. For Intel Foundry, TSMC’s momentum sharpens competitive pressure. TSMC’s scale, lead in advanced nodes and packaging, and ability to align capacity with hyperscaler demand make it hard for a late entrant to match short-term supply and ecosystem breadth. Intel will need to differentiate through competitive node performance, aggressive packaging/assembly offerings, strong design enablement, and regional capacity commitments to win AI accelerator business. The practical implications for the AI community are clear: chip availability, cost-per-inference/training, and deployment timelines will be shaped by how quickly either foundry can deliver next-gen silicon and integrated packages that meet the massive, growing needs of large-scale model training and inference.