Intelligent automation is emerging as the engine of enterprise modernization, as artificial intelligence agents shift from experimental tools to core components of infrastructure and software strategy.

More recently, tech giants such as Nvidia Corp., Salesforce Inc., Dell Technologies Inc. and Intel Corp. have accelerated this transformation through high-stakes acquisitions, platform redesigns and hardware roadmaps built for reasoning at scale. What’s unfolding is not an incremental upgrade but a structural overhaul — where orchestration, inference and agentic systems are redefining how modern stacks are built, scaled and governed, according to John Furrier (pictured, left), executive analyst at theCUBE Research. He unpacked this shift in a thought-provoking discussion with Dave Vellante (right), chief analyst at theCUBE Research, where the two explored what’s next for AI infrastructure during the latest theCUBE Pod episode.

“Sovereign cloud basically is an in-country on-premise deployment or on-country cloud telco,” Furrier said. “It has a lot of similarities as on-premises. On-country means you’re in the country. It’s like on-premises. If you’re on-country, that’s basically private cloud. It’s all hybrid. It’s a data center. The country is the computer. I think sovereign cloud will be the hottest market in the next two years.”

Intelligent automation transforms the software stack

Intelligent automation is rewriting the software playbook. Software-as-a-services apps are giving way to agents that make decisions, take action and integrate behind the scenes. Powered by retrieval-augmented generation, reasoning models and orchestration, this shift is forcing vendors to rethink everything — just look at Salesforce’s $8B Informatica bet on an agent-first future, according to Vellante.

“I don’t think SaaS is going to go away,” Vellante said. “I think agents don’t replace SaaS … agents orchestrate SaaS and become the high value layer on top of that. I do think that the value creation of agentic will be definitely 10x the SaaS business, possibly your 100x. But I think that the business logic that exists inside of SaaS applications that won’t just go away. We might see SaaS get relegated to a lot lower layer of the value chain.”

Still, success will depend on executing across multiple layers — not just AI models, but data harmonization, security, compliance and developer accessibility. Salesforce’s massive Trailblazer community, cultivated over two decades, may become a critical asset in realizing that future, Furrier explained.

“They have to nail the data integration, they got to nail the data cloud, they got to bring in this next intelligence wave. We are on the ‘do the search thing with RAG, get some reasoning going, and then the third wave is intelligence,’” Furrier said. “That’s going to be autonomous, semi-autonomous actions, agents, multi-step agents, agents of agents and then that intelligence where the value will scale up. If they can get their AI capabilities built by then, they could be the AI software cloud.”

Nvidia fuels the infrastructure behind automation

As software shifts to agents, Nvidia is locking in its role as the AI infrastructure backbone. With inference demand exploding — Microsoft Corp. alone hit 100 trillion tokens last quarter — its Grace Blackwell chips and Spectrum-X networking aren’t just upgrades; they’re powering a full data center redesign built for the age of intelligent automation, according to Vellante.

“The data suggests that if you pull out of China and you add in networking, the upside more than backfills for the gap and the gross margin trajectory is there, the share buyback pace is there,” he said. “The bottom line is Blackwell is paying for itself in real time and inference is compounding that effect. The only thing is the export regime looms, and that’s China revenue. Otherwise, it’s all systems go.”

Nvidia’s tightly integrated stack — compute, memory and networking — powers the AI factory: infrastructure built for speed, scale and nonstop reasoning. It’s a bold shift from general-purpose models and a big challenge to rivals. Some say Nvidia’s real moat isn’t exclusion — it’s enabling customers to scale faster with everything built in, Furrier noted.

“The question is, the moat might not be a moat; it might be an enablement value where if Nvidia can throw off the enablement at scale with the tokens and feed the demand and be the systems for that — the data center is the computer, and they’ll roll right into sovereign cloud,” he said.

The new rules of software and scale

While Nvidia leads the GPU charge, Dell and Intel are staking their ground in the intelligent automation boom. Dell’s message is clear: Enterprises want flexibility. With thousands of GPUs already deployed in AI Factory setups, Dell is offering not just hardware, but full-stack systems tailored for sovereignty and scale, Vellante explained.

“It’s just going to take focus in stitching that together. To the extent that you can go, if you can go to organizations with a solution … you bring solutions to the market and you do that potentially by vertical industry through your VARs and other partners,” he added. “I think the cloud will continue to grow faster, but I do think that on-prem will modernize to the extent that these companies, like Dell, like HPE, like IBM, like Oracle … to the extent that they build out their on-prem AI stacks and ecosystems, I think it’ll get significant uptake.”

Modern computing is being rebuilt from the inside out. Agents are becoming operational layers, data drives split-second reasoning and infrastructure is now orchestration and acceleration. This is intelligent automation — and like past tech revolutions, it will reward those who move fast and rethink everything, while leaving legacy thinkers scrambling to catch up.

“I think we’re going to enter an era of urgency around people’s careers, and I think AI’s going to only enable that. I’m very bullish at the fact that it’s going to transform jobs from a more higher knowledge worker,” Furrier said. “I came into the workforce and people didn’t even know how to use a PC. I watched really smart people … become irrelevant really fast. They were smart, but they would lean on their old tools, which is paper and pen and slide rule and all those things. I think we are in that generational shift.”

Watch the full podcast below to find out why these industry pros were mentioned:

Brian J. Baumann, founder of NYSE Wired and director of capital markets, technology at NYSE
Bill Tai, venture capitalist, athlete, adjunct professor
Richard Branson, founder of Virgin Group
Jensen Huang, founder and CEO of Nvidia
David Floyer, analyst emeritus at theCUBE Research
Michael Dell, founder, chairman and CEO of Dell Technologies
Matt Garman, CEO of AWS
Mai-Lan Tomsen Bukovec, vice president of technology at AWS
Satya Nadella, chairman and CEO of Microsoft
Marc Benioff, chair and CEO of Salesforce
George Gilbert, principal analyst at theCUBE Research
Mary Meeker, general partner at BOND
Mark Castleman, managing director for AI cloud at Intel
Katie Potter, manager UX content (GenAI and CMSs) at Extreme Networks
Vincent Molinari, founder and CEO of FINTECH.tv
Paul Martino, founder of Bullpen Capital

Here’s the full episode of this week’s theCUBE Pod:

Don’t miss out on the latest episodes of “theCUBE Pod.” Join us by subscribing to our RSS feed. You can also listen to us on Apple Podcasts or on Spotify. And for those who prefer to watch, check out our YouTube playlist. Tune in now, and be part of the ongoing conversation.

Photo: SiliconANGLE