Software

AI + ML

‘Undigestible rate of change’ sees vendor supergroup create ‘The Open Compute of AI’

It’s called ‘Cosmos’ and Nvidia, Cisco, X, SuperMicro and VAST Data all think it will help – them and you


A group of top enterprise vendors feel that AI is changing so fast it’s “undigestible” to many, so they’ve created an org they hope will function like the Open Compute Project so you don’t make a meal of it.

The “undigestible rate of change” quote was offered to The Register by VAST Data co-founder Jeff Denworth, who told us: “The pace of evolution of AI is nothing short of breathtaking.” Users who try to deploy AI, he said, find it “hard to keep track.”

Denworth lamented the fact that vendors each define their own approaches to AI, leaving the industry without a neutral “watercooler” at which to share experiences or resources.

VAST Data founded an org called Cosmos to become that watercooler. Denworth said one model for the org is the Open Compute Project, the outfit that designs hardware and best practices derived from member organizations’ own efforts and open sources them for anyone to share. Another model is the RSA conference that, despite bearing a vendor’s name, has become a security event that touches on any and all infosec topics.

Jonas Rosland, who has filled community manager roles at VMware, enterprise Linux outfit CIQ, and Dell EMC, has been hired to make Cosmos work. Denworth said a code of conduct has been developed and recruitment of experts to lead content tracks has commenced.

As an example of how Cosmos might work, Denworth told The Register an org working on large language models has been approached to manage an ethical AI track.

Speaking for VAST, he said the vendor will “happily prioritize” features that track develops.

Denworth admitted Cosmos is in a chicken-and-egg moment because it’s yet to produce the content that proves its worth. But he is willing to wait. “This is not a sprint,” he said, adding that the org has sufficient resources to have begun planning for events around the world.

Plenty of the group’s current membership roster - NVIDIA, xAI, Supermicro, Deloitte, WWT, Cisco, CoreWeave, Core42, NEA, Impetus, Run:AI, and Dremio – have decently deep pockets. Nvidia boss Jensen Huang helped to launch Cosmos, which won’t hurt its credibility.

Its structure might: Open Compute was founded around Facebook’s hardware designs, which it shared to enlarge the ecosystem of possible suppliers and out of altruism. The members of Cosmos have something to sell you. And the org’s full name is “Cosmos by Vast.”

Indeed, on the same day as it revealed Cosmos VAST Data announced new collaborations with Nvidia, Cisco, and Equinix that our sibling site Blocks and Files has explained here and here. ®

Send us news
3 Comments

Nvidia won the AI training race, but inference is still anyone's game

When it's all abstracted by an API endpoint, do you even care what's behind the curtain?

AI bubble? What AI bubble? Datacenter investors all in despite whispers of a pop

Billions continue to pour into bit barns across the globe

AI running out of juice despite Microsoft's hard squeezing

Biz leaders still dream of obedient agents replacing workers. In the actual workplace, they're going AWOL

Google teases AI Mode for search, giving Gemini total control over your results

It's just an opt-in Labs curio for now, but so were those ever-present Overviews

CoreWeave rides AI wave with IPO filing – but its fate hinges on Microsoft

GPU farm discloses 77% of revenue tied to just two customers, putting Redmond giant front and center

Cash torrent pouring into Nvidia slows – despite booming Blackwell adoption

May we all have problems like annual revenue growth dropping from 126 to 114 percent

Surprise! People don't want AI deciding who gets a kidney transplant and who dies or endures years of misery

Researchers find AI isn't ready to help with moral decision making

Ex-Googler Schmidt warns US: Try an AI 'Manhattan Project' and get MAIM'd

That's Mutual Assured AI Malfunction in the race for superintelligence

It begins: Pentagon to give AI agents a role in decision making, ops planning

Former allies, take note. This is a WOPR

AMD looks to undercut Nvidia, win gamers' hearts with RX 9070 series

The question is whether we can find them in stock and at MSRP

Despite Wall Street jitters, AI hopefuls keep spending billions on AI infrastructure

Sunk cost fallacy? No, I just need a little more cash for this AGI thing I’ve been working on

Dems ask federal agencies for reassurance DOGE isn't feeding data into AI willy-nilly

Pouring sensitive info into unapproved, unaccountable, unsafe models would be a 'severe' cybersecurity fail