Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Marvell reveals optical shared-memory module for AI servers

AI summary

Marvell revealed its Photonic Fabric Memory Module, an optical-interconnect design intended to place shared memory away from processors and link it across AI servers. Sixteen modules are planned to form an appliance with 32TB of DDR5 and 1,152GB of HBM3e shared across as many as 16 hosts; each module is designed with 72GB of HBM3e, up to eight DDR5 DIMM slots, a Photonic Fabric ASIC and 7.2Tbps of optical I/O. The design remains pre-silicon, has no measured throughput on working hardware and is targeted for customer sampling in 2027.

In-depth insight

Marvell says the shared-memory tier could support cross-node KV caches and allow memory capacity to scale independently of accelerators, although TechRadar reports that cost savings and performance remain unquantified.

  • Marvell reveals Photonic Fabric Memory Module, the building block of a 16-module appliance sharing 32TB of DDR5 and 1,152GB of HBM3e across 16 hosts over fiber
  • The design dates back to Celestial AI's 2025 presentations, currently remains pre-silicon, and is targeting customer sampling by 2027
  • Nvidia is already hedging with a $2 billion Marvell stake in addition to its own BlueField-4 STX context memory, Enfabrica's pooling tech, and a co-packaged optics roadmap

Marvell has put physical hardware behind one of the more ambitious ideas in AI infrastructure.

In a video posted to X, and earlier in a theCUBE interview recorded around its October 6 Investor Day, the company showed off its Photonic Fabric Memory Module, which it bills as the "first optically interconnected high-capacity, high-bandwidth memory."

The pitch is simple: use light instead of copper so that dozens of terabytes of memory can sit in a separate box, racks away, and be shared by many processors at once, something that could effectively birth the widespread use of communal or shared memory across multiple AI servers.

Two terabytes at a time

Some clarity, however, is warranted: the 32TB figure belongs not to the module but to the Photonic Fabric Memory Appliance, which gangs 16 modules together and shares the pool across up to 16 hosts.

#Marvell#Photonic Fabric Memory Module#Photonic Fabric Memory Appliance#Photonic Fabric ASIC#Celestial AI#Nvidia#BlueField-4 STX#Enfabrica#EMFASYS#NVLink Fusion

Original · 5 min readRead the original
Group chat

No comments yet. Start the conversation.