Samsung foundry targets robotics, auto AI chips with Cadence platform - digitimes

Samsung Electronics is expanding its foundry business into physical AI semiconductor markets through a chiplet platform developed with Cadence, targeting applications in robotics, automotive systems, ... Read original →

Samsung’s push into robotics and automotive AI chips via Cadence’s chiplet platform is a direct challenge to TSMC’s dominance in HPC and automotive electronics. Technically, this move pressures the entire stack—from EDA tools to advanced packaging like FOPLP—accelerating supply chain upgrades, especially benefiting OSATs in Taiwan, China and Korea. On compliance, tightening U.S. export controls on advanced equipment could inflate Samsung’s non-U.S. fab certification costs; heavy reliance on U.S.-origin Cadence IP may also hinder its traction with Chinese automakers due to geopolitical friction. Strategically, TSMC will likely fast-track CoWoS-L capacity and deepen Synopsys ties, while UMC doubles down on mature-node automotive MCUs. Over the next 18 months, delayed chiplet standardization will fragment customer designs, paradoxically strengthening ecosystem lock-in by leading foundries—making Samsung’s play less about volume and more about capturing architectural authority in the AIoT era.