Most Enterprises Are Spending on AI and Hoping. That's the Real Story.

The AI Investment Paradox

Here's something that should make every technology leader uncomfortable: according to KPMG's Global AI Pulse Survey, three out of four global leaders will prioritize AI investment despite economic uncertainty. And 65% of UK respondents say their organization would continue investing in AI regardless of tangible ROI.

Read that again. Not "even if ROI is slow." Regardless of it.

I've seen this pattern before. In my experience, when companies start treating a technology investment as a cost of doing business rather than something to be justified, it means one of two things: either the technology has genuinely become table stakes, or fear is driving the bus. With AI in 2026, I believe it's a bit of both.

A Performance Gap, Not Just a Maturity Gap

KPMG's findings surface a distinction that matters a great deal. Among organizations leading in AI adoption, 82% report that AI is already delivering meaningful business value. Among their peers? Only 62%.

That 20-point gap is not about who has more GPUs or a bigger innovation lab. KPMG puts it plainly: this is a widening performance gap between organizations that treat AI as enterprise-wide transformation and those bolting AI onto existing models for incremental gains.

I see this constantly in advisory work. A company launches ten AI proofs of concept across different departments, each one siloed, each one with its own data pipeline and success criteria. Then they wonder why nothing scales. The issue isn't the AI. It's that the organization never did the harder work of rethinking how it operates.

Why Traditional ROI Doesn't Work for AI

So why is ROI so elusive? The article from CIO.com gathers analyst perspectives that paint a nuanced picture. I think there are at least four distinct problems happening at once:

• We never measured the baseline. AI is replacing intellectual effort that was never quantified in the first place. How do you calculate savings on something you never tracked? As Ben Grant from Lambton Capital Partners put it: "The value shows up in time reclaimed, decisions made faster, and gaps being plugged before they become problems. Try putting that in a spreadsheet."

• PoCs are set up to fail. Executives launch AI proofs of concept with unrealistic ROI goals. When the LLM can't deliver against a standard that was technologically impossible to begin with, the project gets labeled a failure. That's a goal-setting problem, not a technology problem.

• Hidden costs keep appearing. One example that caught my eye: enterprises deploying customer chatbots are discovering that users abuse them as free genAI tools, and the company ends up paying for all those extra tokens. The cost model for AI is genuinely different from traditional software, and many organizations haven't internalized that yet.

• The old ROI playbook doesn't fit. Michael Leone from Moor Insights & Strategy nails this one. CIOs can tell you the productivity gains on a specific workflow, but ask about the three-year enterprise payoff and you get a shrug. AI isn't ERP. It doesn't follow the same deployment and value realization curve.

The FOMO Factor Is Real

Let's be honest about what's driving a lot of this spending. It's not all strategic vision. A significant portion is competitive anxiety.

Manish Jain from Info-Tech Research Group frames it well: "When a new engine comes along, wise operators don't ask first what it earns. They ask what happens if they're the only ones without it." And independent analyst Carmi Levy goes further, calling investment without ROI "sheer fiscal suicide" in principle, but acknowledging that "AI now compels organizations to dive in more out of fear of being left behind."

I find this tension fascinating and familiar. Boards are telling CIOs that AI investment is not optional. But the same boards will eventually ask where the returns are. The clock is ticking on that contradiction.

When AI Becomes the Office Suite

There's a perspective from Gartner's Nader Henein that I think deserves more attention. He argues that some AI investments, like AI assistants, are becoming standard office tools. "No one calculates ROI by counting the number of Word documents or presentations produced," he says.

This is an important framing shift. But there's always a but.

Henein also warns that if an AI investment "burns cash and fails to produce any tangible ROI, it will be retired. P&L reports and the expectations of investors from publicly traded companies are not changing." That's the reality check. AI may be the new normal, but it still has to justify its existence on the balance sheet eventually.

The Uncomfortable Truth: Maybe 1 in 10

The quote that stuck with me most from the original reporting was from Michael Leone:

"Maybe one in ten enterprises I've spoken to has the talent, governance, and operating discipline to actually get compounding returns from its AI spend. Everyone else is spending and hoping. That's the real story."

One in ten. That is a sobering number. And honestly, it aligns with what I see in my own work. The organizations pulling ahead aren't the ones with the biggest budgets. They're the ones with clear governance, the right people, and the discipline to treat AI as an operating model shift rather than a technology bolt-on.

What This Means for Technology Leaders

If you're a CTO, VP of Engineering, or a senior architect reading this, here's what I'd take away:

1. Stop using old measurement frameworks for new technology. If the work AI replaces was never measured, you need new metrics. Time reclaimed, decision velocity, error reduction, capacity freed for higher-value work. Build those measurement capabilities now, before the CFO comes asking.

2. Set honest PoC goals. If you let the business set AI targets that are technologically unreachable, you're setting up the team and the technology to fail. Push back on fantasy metrics.

3. Model for hidden costs. Token costs from unexpected usage patterns, data pipeline complexity, security and privacy overhead. These are real and they're catching teams off guard.

4. Invest in governance and talent, not just models. Budget and mandate are no longer the blockers. Security, privacy, and the ability to run AI at scale are. If you're in the nine out of ten that lack the operating discipline for compounding returns, that's where your attention should go.

5. Don't confuse strategic patience with strategic blindness. KPMG's Leanne Allen said it well: viewing AI as a long-term investment is an important mindset shift, but "that shouldn't translate into investing in AI blindly, without a clear strategy." The fact that ROI is hard to measure doesn't mean you shouldn't try.

The gap between AI leaders and the rest is widening. And it's not widening because of technology. It's widening because of how organizations think about, govern, and operationalize AI. If your company is "spending and hoping," the question isn't whether to keep investing. It's whether you're building the foundations that will eventually turn that spend into compounding returns, or just burning cash while the competition figures it out first.