The way AI is being talked about today — especially in boardrooms, investor decks, and venture capital echo chambers — has stopped being thoughtful and started being hysterical. And candidly, in my day-to-day work running a fintech company, the hysteria simply doesn’t square with the reality.

I’ve lived through all of it: the internet, mobile, crypto, blockchain, cloud computing. All of them mattered. None of them mattered in the same way — and that distinction is exactly what today’s AI conversation keeps getting wrong.

The question I find most clarifying is this: “For a given company or business: Is AI more like the internet — or more like cloud computing?”

The internet changed how we live, communicate, and do business. Companies had to reinvent themselves to survive. Cloud computing was also a massive technological shift, one that created trillions of dollars in market capitalization for its providers. But for many operating businesses, it didn’t radically change what they did, — just how efficiently they did it.

I’ve run companies on both sides of that transition. Before the cloud, we rented space at hosting centers, bought hardware, and ran 24/7 network operating centers. It worked. When I started Capitolis, the cloud was mature and available. It made scaling easier and more cost-efficient, but the underlying business model didn’t change. That kind of decision is rarely a CEO or board-level call. It’s an engineering leader working with a CFO and COO: What’s the return? What’s the trade-off?

That’s exactly how many companies should be approaching AI right now.

The AI Reality Check

There are businesses that will be completely transformed — or destroyed — by AI. I recently had a customer service interaction with DoorDash that seemed AI-driven. It was fast, accurate, and better than what most human-staffed help desks deliver. For companies like that, the hype may actually be understated. But this isn’t universal.

At Capitolis, we are a highly integrated B2B network embedded within the institutional financial world. We are not going to be disrupted by AI any time soon. We are investing in AI — hundreds of thousands of dollars a year — and implementing it throughout the organization. We see pockets of efficiency. But we have not seen the returns on that investment yet.

In engineering, we see agents writing code and believe we can deliver something like a 25% productivity gain over time. With roughly 100 developers, that’s meaningful. But the tough questions still apply: Where does this rank against competing priorities? How long will it take, and at what cost? What gets delayed?

When you look at actual ROI, the picture is sobering. Much of what many companies are spending on AI right now is exploratory, not transformative. That’s fine — as long as we’re honest about it.

The Noise Is the Problem

If you listen to the loudest voices in the room, you’d think every company is on the brink of existential failure if it doesn’t reorganize itself around AI immediately. Every conference, every pitch, every panel — it’s the only conversation anyone seems capable of having. That framing is wrong and potentially damaging.

Some AI-native companies hold genuine promise and will reward their investors. But that doesn’t mean every operating business should behave as if AI is an all-consuming threat. For many healthy, fast-growing, profitable businesses, AI will look much more like cloud computing: a powerful efficiency tool, not a mandate to rewrite the business model.

Make independent, ROI-based decisions. Ignore the noise. Treating AI like an all-consuming existential threat is not a strategy. What we need is less hype and more discipline. AI deserves seriousness. The hysteria doesn’t.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.