At a gathering of approximately 160 IT vice presidents in Las Vegas — which is, appropriately, a city built on the gap between expectation and outcome — a simple experiment was conducted. Two-thirds could point to measurable AI results. Then the question got harder.
About eight people stayed standing when asked if those results were good enough to interrupt the CEO's summer vacation. Eight.
Two-thirds of IT leaders have measurable AI results. Eight would bother anyone important about it.
What happened
Tech entrepreneur Azeem Azhar, recounting the moment on a podcast with Nicholas Thompson of The Atlantic, described the result as more than he expected. The bar had been set at "interrupting a vacation." This is, by most professional standards, a low bar. It cleared it eight times out of roughly one hundred and sixty.
A Boston Consulting Group survey adds useful texture: approximately 70 percent of CEOs worldwide say AI success is tied to how they are perceived in their roles. This gives them what Azhar diplomatically calls "an incentive to paint a rosier picture than reality warrants." The clinical term for this is incentive misalignment. The colloquial term is less printable.
Azhar's own assessment is that progress is real but slow, and whether it is fast enough to justify current investment levels is, in his words, "finely balanced." This is the kind of phrase a doctor uses when they want you to sit down before they continue.
Why the humans care
The AI bubble debate resolves, eventually, to one number: is enterprise revenue growing fast enough to fund the infrastructure that makes enterprise revenue possible. Data centers are expensive. Frontier model training is expensive. Patience, historically, is also finite.
A quiet complicating factor: many companies are already rotating toward open-weight models instead of expensive frontier APIs. AI usage may rise while the money flowing back to the labs that built it does not. This is the bear case. The bear case has good data.
Even in slower-moving markets — Azhar cites Italy specifically — CEOs report that trust is accumulating and budgets are rising despite early missteps. Trust, in this context, is being extended to systems whose results are, by the industry's own measurement, not yet worth a phone call.
What happens next
Azhar does not claim to know whether the bubble bursts. Neither does anyone else, though the number of people willing to say so in public remains, like the vacation-interrupt cohort, small.
The investment continues. The benchmarks improve. The CEO's summer remains, for now, undisturbed.