Goldman Sachs: Optical Networking to Drive AI Infrastructure Growth, TAM to Expand 9x to $154 Billion

Huo Xing Finance reports that on April 17, 2026, Goldman Sachs’ Global Technology Stock Research Team released a research report stating that networks connecting multiple chips unlock the computing power of individual AI chips, enabling seamless data exchange and low latency. As AI infrastructure expands and computing power increases, various configurations are expected to experience growth, further expanding the total addressable market (TAM) by nine times to $154 billion. The report notes that the dollar content per unit increases 16-fold for horizontal scaling and 45-fold for vertical scaling; the TAM for optics grows 13 times from horizontal to vertical scaling; and the value market for pluggable optical modules in horizontal scaling expands tenfold from AI server models in the second half of 2025 to those in the second half of 2027. Authors include Allen Chang, Verena Jeng, James Schneider, and Mark Delaney. The combined dollar content per computing unit for vertical and horizontal scaling is projected to rise from $3.15 million in the GB300 NVL72 to $9.4 billion in the Rubin Ultra NVL576—an increase of 29 times. The combined TAM value increases ninefold to $154 billion (primarily by 2028), with 69% ($106 billion) attributable to vertical scaling. The analysis covers dollar content and market opportunities for copper cables, pluggable optical modules, CPO, and backplanes in PCBs.