TSMC: Physical AI Drives Chipmakers Up the Value Chain

At the TSMC European Symposium, semiconductor CEOs revealed how physical AI—AI integrated into robots and autonomous systems—is pushing chipmakers to deliver more integrated solutions, moving beyond traditional chip supply into system-level value, impacting substrate and PCB design complexity.

The focus on physical AI, which includes applications like humanoid robots and autonomous vehicles, demands higher performance per watt and tighter integration. TSMC executives noted that this trend forces chipmakers to collaborate earlier with packaging and substrate partners. For PCB manufacturers, this means increased demand for advanced HDI, IC substrates, and multilayer designs that support heterogeneous integration and high-speed signaling. The shift also accelerates adoption of glass and ceramic substrates for thermal management and signal integrity in AI compute nodes.

European chip CEOs at the symposium stressed that physical AI requires moving up the value chain from component to solution provider. This includes offering complete subsystems with embedded PCBs and substrates optimized for AI inference at the edge. LT CIRCUIT, a manufacturer specializing in HDI, multilayer, and IC substrate PCBs, sees this as a direct driver for more complex, high-layer-count boards and rigid-flex designs. The company notes that customers are already requesting substrates capable of supporting chiplets and high-bandwidth memory for AI accelerators.

LT CIRCUIT, a leading PCB/PCBA manufacturer, offers advanced HDI, multilayer, Rogers, rigid-flex, ceramic, and IC substrate PCBs. The company sees rising demand for complex board designs driven by physical AI and heterogeneous integration trends in chip packaging.