Nvidia · AMD · Elon Musk · SpaceX · Tom's Hardware
Inside optical and the battle for scale, how the AI industry is racing to integrate photonic interconnects
Compiled by KHAO Editorial — aggregated from 1 source + 1 reference discovered via search. See llms.txt for citation guidance.
◌ Single Source
“Moore's Law is deader than a doornail, absolutely toast,” said Nick Harris, chief executive of photonics firm Lightmatter, 's Hardware Premium.
Key facts
- She guesses that they coexist, split by application, with a rough 60/40 lean towards CPO
- I think 2027 and 2028 are going to be very, big years for near-package optics,” Harris said
- Engineering samples are due in the second half of 2026, with mass production slipping to Q2 2027, while AMD's next-generation MI500-series rack is expected to follow quickly
- Lightmatter is already building CPO parts at TSMC and GlobalFoundries, which it expects to ship in 2028
Summary
AI hardware has been a pretty simple game for years: Whose GPU has the most FLOPS, the most memory, or the fastest clock? It all comes down to scale-up and scale-out. As Polina Bayvel, professor of optical communications and networks at University College London, told them, the hyperscalers are "on a drug which says that the more you put in, the better is the answer", and between 72 and 144 GPUs now go into a single scale-up rack before it needs to reach outward. Copper was sufficient within racks for years, but it has hit a wall, and a supply shortage. Today, optical signals usually arrive through pluggable transceivers, the standard modules that sit around 18 inches from the switch. After that comes co-packaged optics, or CPO, where optical chiplets sit around the GPU or switch.