The AI industry has spent the last couple of years obsessing over models. Bigger models, faster models, new models every few months.
Meanwhile, many companies are still trying to answer a more practical question… where exactly is all this stuff supposed to run?
Equinix partners with Cisco and NVIDIA on AI Factory
That’s part of what sits behind a new announcement from Equinix, which is expanding its work with Cisco and NVIDIA to help customers deploy the Cisco Secure AI Factory with NVIDIA across its global data center footprint.
The company says customers will have access to standardized deployment blueprints and automation designed to make AI infrastructure easier to stand up.
Presidio’s P.A.T.H. Lab gives customers a testing environment
Equinix is also working with Presidio on something a little more hands-on: a new testing environment called the P.A.T.H. Lab, where customers can try out AI infrastructure before rolling it into production.
“The success of enterprise AI starts with its physical foundation,” said Gordon Mackintosh, senior vice president, global partner sales and ecosystems, Equinix. “Our collaboration with Cisco, NVIDIA and Presidio delivers the infrastructure AI workloads demand while giving customers a place to prove it out before they scale.”
That “prove it out” piece may be the more interesting part of the announcement.
One theme that keeps showing up in enterprise AI is that getting a pilot running isn’t necessarily the hard part anymore – plenty of organizations have dabbled in the AI arts.
The more fun challenge is figuring out what happens next. Where does the data live? How much infrastructure is actually needed? What happens when a project moves beyond a small group of users? These are the questions.
Enter, P.A.T.H. Lab. Built on Cisco’s Secure AI Factory with NVIDIA, the environment is designed to provide customers with a place to test, validate, and refine AI deployments before making larger investments.
Enterprise AI infrastructure moves beyond pilot projects
Presidio sees that as a response to a different conversation happening in the market today.
“One of the most important shifts we’ve seen in the last 18 months is that AI success is no longer about finding the most powerful model,” said Tim McHugh, VP partnerships & alliances, Presidio. “It’s about building the infrastructure that can run AI everywhere it matters, without sacrificing data sovereignty or control.”
This is another sign that the industry is spending more time talking about infrastructure, deployment, and operations.
The excitement phase isn’t over, but a growing number of vendors seem focused on the less glamorous work of helping customers turn AI projects into something they can actually run day after day.
Infrastructure grabs the mic
The announcement itself is about infrastructure, but the P.A.T.H. Lab may tell a different tale about where the market is headed. Customers have heard plenty (and we mean plenty) of AI promises by now.
More and more, they want a chance to test the technology in a real environment and see what it actually takes to run before making larger investments.
Cisco and NVIDIA have spent much of the past year talking about what happens after the AI proof of concept. Their Secure AI Factory initiative is built around many of the same themes as the Equinix announcement, including infrastructure, security, and a clearer path from experimentation to production for organizations.