LAS VEGAS -- Predictions that governance concerns, data gravity and more accessible AI models would prompt enterprise private AI growth on-premises are showing signs of coming true.

Speakers and attendees at VMware Explore this week shared their experiences running open-weight and open source AI models on-premises and developing AI agents, along with some shifts in thinking about the feasibility of private AI.

Broadcom has taken an opinionated stance on the advantages of private cloud over public cloud services since it acquired VMware in 2023, and officials continue to maintain that most enterprise AI workloads will eventually return on-premises.

While this has yet to materialize, one attendee said his stance on that message has shifted since Broadcom CEO Hock Tan's 2024 keynote extolling private cloud.

"It was funny when Hock Tan did his keynote. He said, 'The future is on prem!', and I'm like, 'Ha ha ha ha, No, it's not.' But here, Hock has the last laugh," said Scott Laforet, VMware product owner at a North American bank he requested not be named due to company policy prohibiting him from representing it in the press. "I'm a service provider under the covers. ... What I don't offer presently is anything with a GPU. But the next [generation] of workloads must accommodate that."

Specialized models + governance = enterprise agent success

The latest generation of GPUs can still be pricey and hard to come by, as can memory amid a global supply shortage, but older generations of GPUs and even CPUs can now accommodate smaller, specialized and open models that have become more powerful over the past two years.

"When we started this, everyone thought we were going to need the best models for this, and in just the time it took us to plan to execute on -- and you've been able to see this across all AI -- all these models are getting so much better every day," said Adam Little, director of IT operations at New Belgium Brewing in Fort Collins, Colo., during a breakout session presentation.

One example is a group of open models built for Nvidia's Omniverse platform, which New Belgium has used to visualize and optimize centrifuges in its brewery manufacturing pipelines. As the company moves to agentic automation for multiple parts of its brewing process and business, many models have now come into play, Little said, including models trained on New Belgium's internal data.

"A frontier

Enterprises make strides with private AI on-premises

Suggested for you

More Stories

Trump quietly begged Army secretary to stay through midterms — he quit anyway: report

1Password Wades Into a Right-Wing Mess After Funding a Linux Project

Freaked-out Republicans get secret Rose Garden invite as Trump clings to massive war chest

iOS 27 Introduces New 'iPhone Handoff' Feature

Thousands Of Dropbox Accounts Breached In 'Verification' Hack

Pink house's new auction date set

Discussion