Jensen Huang: Nvidia's Agentic AI Will Make Its Biggest Move Yet!

Added: 2026-05-11

[00:00:00]in Las Vegas. The both shares are fractionally lower, John, but with more on what does sounds like a very big announcement. Jensen Huang is making Nvidia's biggest move yet.

[00:00:08]And he's doing it with the CEO of ServiceNow live on stage in front of the entire enterprise software world.

[00:00:14]>> This is one of the greatest transformations for the software industry ever. For for the first time, service is software. Software is service. Nvidia announced a new autonomous agent called Project Arc. It can code, think, and work with enterprise applications all on its own.

[00:00:34]What does this mean for Nvidia's stock?

[00:00:36]Why does this unlock a market 100 times larger? And why could the ripple effect be bigger than anything we've seen from this company yet? Here's what's actually happening. Nvidia didn't just build another artificial intelligence model.

[00:00:49]Project Arc is an autonomous agent, [music] and that word, autonomous, is doing a lot of work here. It means the agent doesn't wait for instructions. It reads a situation, decides what to do, and takes action entirely on its own. It can write code, navigate enterprise software, and complete complex multi-step tasks without a human touching it at any point. Think of it like hiring a highly capable employee who never sleeps, never gets distracted, and can work across every department at the same time.

[00:01:19]That's the direction this technology is heading, and it's arriving faster than most companies are prepared for.

[00:01:24]But here's what makes the ServiceNow partnership specifically important.

[00:01:28]ServiceNow is the operating layer that already lives inside most large companies. It manages IT requests, approvals, workflows, compliance reporting, the unglamorous but absolutely essential infrastructure that keeps businesses functioning every single day.

[00:01:46]When Nvidia builds an autonomous agent that plugs directly into that infrastructure, this stops being a tech demo. It becomes a product that every major enterprise on the planet could actually deploy.

[00:01:57]>> We have to be the agentic front door. We work with great companies like Nvidia, the great Jensen, the great Nvidia. Uh we announced project Arc. You'll hear you'll hear about that today. I'm sure Jensen will have some thoughts on it.

[00:02:09]But we want to manage the identities of the humans and the agents. We treat the agents just like the humans. Right now, most enterprise artificial intelligence deployments fail at the same point. Not because the technology doesn't work, but because companies can't control it. Who does the agent report to? What systems can it access? What happens when it makes a mistake and there's a compliance violation? ServiceNow's answer to that problem is what makes project Arc deployable in the real world, not just inside a research lab, but inside actual companies that have regulators, auditors, and legal obligations to satisfy.

[00:02:48]Now, let's talk about the business side, because the numbers here are genuinely hard to ignore.

[00:02:54]ServiceNow entered 2019 as a $3.5 billion company. This year they're projecting nearly 16 billion in revenue, and they just told analysts they plan to double the company again by 2030, putting them north of 30 billion. This year we're projecting to be nearly a 16 billion company.

[00:03:13]Right now, we have a 28 billion RPO. So, we're growing faster than any other enterprise software company at scale in the world ever ever happened. Faster growth than any enterprise software company at this scale in history, and the bet is that agentic artificial intelligence is what drives ServiceNow from 16 billion to 30 billion.

[00:03:35]Here's why that target is credible rather than just ambitious. ServiceNow already has 28 [music] billion dollars in committed future revenue sitting on its books right now.

[00:03:46]These are contracts that are already signed, but not yet delivered. So, this isn't speculation about future demand.

[00:03:52]The pipeline is already locked in. The question is purely how fast it compounds from here and how much the agentic shift accelerates that timeline.

[00:04:01]The Nvidia partnership is central to that acceleration. ServiceNow doesn't need artificial intelligence that generates polished reports. It needs artificial intelligence that can actually execute inside complex business workflows.

[00:04:14]File the compliance form, escalate the security incident, process the employee request. That is precisely what Project Arc is built to do.

[00:04:22]Remember what Jensen said at the start?

[00:04:25]One of the greatest transformations in software history.

[00:04:28]We're about to see exactly why he believes that. And the number he cited is one that should make every investor in this space stop and recalibrate.

[00:04:36]Jensen's core argument is straightforward once you strip away the technical language.

[00:04:41]The service industry, think consulting, IT support, customer service, legal processing, finance operations, human resources, is roughly 100 times larger than the software industry. And until now, software couldn't actually perform services. It could support them, automate pieces of them, organize the workflow around them, but it couldn't do the work itself. Agentic artificial intelligence changes that equation entirely. And the service industry is 100 times larger than the software industry.

[00:05:12]And so now, for the very first time, you're going to have human agents that are of course managed and supported by ServiceNow, augmented with AI agents that are going to be working autonomously. That's not a rounding error. That's a completely different addressable market than anything software companies have ever gone after before.

[00:05:35]If artificial intelligence agents can now perform actual service work, resolving an IT ticket, onboarding a new employee, processing a claim, completing a compliance audit, then the companies building the infrastructure for that work become worth an enormous amount. Not because of what they sell today, but because of what the total market becomes once agents replace or augment human service work at scale. And that creates a governance problem that almost nobody has solved yet.

[00:06:04]>> And they're all going to need identity management, access control, network control, all of the things that companies do with respect to regulation and compliance and governance, all of that has to be provisioned whether you're human or an agent.

[00:06:22]>> Every rule that applies to a human employee has to apply to an artificial intelligence agent, too.

[00:06:27]And right now, most companies have no infrastructure to enforce that.

[00:06:32]That is the exact gap ServiceNow and Nvidia are building into together. This is also why the timing creates a real competitive moat.

[00:06:40]We're still in the very early stages of agentic artificial intelligence being deployed inside enterprise environments.

[00:06:48]The companies that build the governance layer now, identity management, access controls, audit trails, compliance frameworks, will become required infrastructure for every business that wants to deploy agents responsibly.

[00:07:01]That's not a feature advantage that a competitor can copy in 6 months. That's structural. Here's where Jensen's numbers get genuinely striking.

[00:07:09]Two years ago, when generative artificial intelligence was just producing text and images, the compute demand was already enormous. Jensen now says that demand has increased by 1,000% just from the shift between generative and agentic artificial intelligence. AI agentic AI needs to be processed in real time. Unlike software of the past that's pre-recorded, you could put it in storage, retrieve it as you need it.

[00:07:34]Today's software needs to be processed completely in real time. Agentic artificial intelligence has to process every single decision in real time, every reasoning step, every tool call, every output.

[00:07:48]That requires a fundamentally different kind of computing infrastructure running continuously, not on demand. Think of it this way. Streaming a pre-recorded video costs almost nothing at the point of delivery, but running a live conversation where the system has to listen, reason, decide, and respond in real time requires active compute the entire time.

[00:08:10]Now, multiply that across millions of agents working simultaneously inside thousands of enterprise systems around the world. The amount of computation necessary from generative AI 2 years ago to now agentic AI, it has gone up a thousand percent. Because the AI now has to read a lot more, use tools, reason, generate a lot of tokens. Agentic systems read more context, use external tools, reason through multi-step problems, and generate far more output than generative artificial intelligence ever did. And they do all of this continuously, in real time, with no off switch. This is the part most market commentary gets completely wrong.

[00:08:50]The popular argument is that smarter software eventually reduces hardware demand, that efficiency gains shrink the need for chips. Jensen is making the exact opposite case, and his logic holds up.

[00:09:02]Smarter artificial intelligence doing more complex work requires exponentially more compute at every step. Every reasoning loop, every tool interaction, every autonomous decision is a compute event. And the hardware running all of those events is Nvidia's. Not only that, the amount of use has gone up orders of magnitude, because now for the first time, AI is doing work. It's doing useful work. For the first time, artificial intelligence is doing actual work, not just answering questions or generating content, but completing tasks that companies used to pay people to do.

[00:09:40]And the company supplying the engine for all of it just showed you exactly where the next chapter leads.

[00:09:45]Project Arc isn't just a product announcement. It's a signal that the compute demand story has a second chapter, and he and Jensen is telling anyone paying attention that this chapter is significantly larger than the first one.