Skip to navigationSkip to main contentSkip to right columnAdam Spatacco, The Motley FoolTue, June 23, 2026 at 5:50 PM GMT+2 4 min readAs graphics processing unit (GPU) clusters grow larger and more power-hungry, electrical connections are reaching their limits when it comes to speed, heat, and energy use. In separate announcements, both Goldman Sachs and Nokia (NYSE: NOK) recently flagged photonics as the next critical layer in artificial intelligence (AI) infrastructure. Optical networking, which uses light to move data, offers an alternative path forward.Let's analyze this enormous commercial opportunity and explore what it could mean for Nokia as the Finnish company quietly transitions from yesterday's leader of mobile devices to an AI networking powerhouse.Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »Image source: The Motley Fool.What is photonics, and why does it matter for AI?Photonics is the process of generating, controlling, and detecting light (photons) to transmit information. In AI data centers, silicon photonics takes this process a step further by integrating lasers, modulators, and detectors directly onto chips using existing semiconductor manufacturing lines. This integration can deliver measurable advantages over traditional electrical wiring, including higher bandwidth, lower latency, and reduced power consumption.As AI training demands ever-larger clusters of GPUs, electrical signals are struggling to keep up without excessive heat and energy waste. Photonics solves this by enabling dense, energy-efficient optical links between chips and racks. This is essential, as compute power is no longer the main constraint of AI development -- connectivity is. Without scalable optical solutions, the next generation of AI applications will be limited by underlying infrastructure rather than silicon's capacity capabilities.How large is the optical networking market?Goldman Sachs calls optical networking the next mega-trend in AI infrastructure. Goldman's analysts project the total addressable market (TAM) for optical networking tied to AI to grow ninefold, rising from roughly $15 billion in 2026 to $154 billion by 2028.The investment bank segments optical networking across two key subcategories. Scale-up networking, which is high-bandwidth connections within racks, accounts for $106 billion of the total TAM. Within scale-up networking are co-packaged optics (CPO), which integrate optical engines directly with processors for maximum efficiency. Goldman estimates CPO to be a $91 billion opportunity at scale.Terms and Privacy PolicyEU DSA contactPrivacy & Cookie SettingsMore Info