Nvidia Mass-Produces 200G Spectrum-X Switch as Networking Revenue Jumps 199% to $15 Billion
Updated
Updated · Yahoo Finance · Aug 15
Nvidia Mass-Produces 200G Spectrum-X Switch as Networking Revenue Jumps 199% to $15 Billion
3 articles · Updated · Yahoo Finance · Aug 15
Summary
Spectrum-X Ethernet Photonics has entered full mass production, making Nvidia the first to ship a 200G-per-lane co-packaged optics Ethernet switch for AI data centers at volume.
The design tackles a key AI scaling bottleneck by integrating optics onto the switch chip, cutting lasers fourfold, lowering network power use 5x and improving signal integrity 64x.
CoreWeave, Lambda and Oracle are the first customers, with production ramping from May through July before reaching full volume.
Nvidia said the launch is already feeding a fast-growing networking business, where revenue surged 199% to $15 billion as it expands control beyond GPUs into data-center infrastructure.
TSMC and Lumentum are supplying the buildout, and Lumentum said its revenue doubled to $1 billion with multi-hundred-million-dollar 2027 orders already secured.
With CoreWeave, Lambda, and Oracle adopting Spectrum-X Photonics, is AI networking becoming Nvidia’s next major growth engine?
If Nvidia’s new photonic switch cuts AI network power 5x, will co-packaged optics become the next must-have technology after GPUs?
Nvidia’s $10 Billion Datacenter Ethernet Surge: How Spectrum-X and Co-Packaged Optics Redefined AI Networking in 2026
Overview
In 2026, Nvidia surged to the top of the datacenter Ethernet switch market, fueled by massive investments from cloud giants building AI infrastructure. This growth was enabled by Nvidia’s acquisition of Mellanox, which provided key networking technology, and the launch of its integrated Spectrum-X platform that combines GPUs, switches, and DPUs. By co-designing the network layer with its GPUs and adopting advanced co-packaged optics, Nvidia achieved major efficiency gains and shifted buyer preferences toward unified, AI-optimized systems. However, Nvidia’s dominance faces risks from supply chain dependencies, manufacturing challenges, and hyperscalers moving to open standards and custom chips.