Breaking the Copper Wall: Co-Packaged Optics and the Materials Race in AI

As AI workloads surge, copper interconnects are hitting hard limits in power and bandwidth density. Co-packaged optics (CPO) solves this by moving optical-to-electrical conversion directly onto the chip package next to switch and compute silicon. This delivers dramatically lower energy per bit and greater bandwidth density—central to NVIDIA, Broadcom, and hyperscalers’ roadmaps. This STEM Talk explores why CPO is vital for AI compute economics and key materials challenges.

Mark Patel, Senior Partner in the Bay Area office, is a leader in McKinsey’s global Semiconductors Practice and a member of the McKinsey Technology Council, focused on serving clients where complex technologies converge for global impact. He has led programs supporting some of the most iconic companies in advanced industries to embrace disruptive technologies across products and operations, combining deep technical understanding with hands-on operational expertise.

Guttorm Aase is a Partner in McKinsey’s New York office and the firm’s global knowledge leader for specialty chemicals. He advises clients across chemicals, materials, and advanced industrials on strategy, M&A, and innovation, with deep expertise in the semiconductor material supply chain. He has led major industry studies on resilient global supply chains and AI-driven R&D transformation, and has advised senior leadership teams across North America, Europe, and Japan for over 18 years. He is a CME Board Member since 2026.