IT Leaders Report AI Results, But Few Show Major Business Value

The core debate surrounding the artificial intelligence bubble focuses on whether revenue at AI labs is growing quickly enough to fund massive data center construction. Key variables include how long purchased AI chips remain useful in production. Four, six, or eight years changes the financial math significantly. The most critical question is whether companies receive enough genuine value from AI to justify continued purchases at rising prices. So far, economy-wide return on investment lacks documentation and appears mainly through individual anecdotes.

British tech entrepreneur and Exponential View founder Azeem Azhar shared a telling anecdote during a podcast with The Atlantic editor Nicholas Thompson. Azhar spoke to roughly 160 IT vice presidents in Las Vegas and asked who could point to measurable AI results. Two-thirds of the room remained standing, which exceeded Azhar's expectations. He then asked who held results strong enough to interrupt a chief executive officer during a summer vacation. Only about eight people kept standing.

Azhar concluded from this exercise that businesses are making progress, but the pace is slow. Observers debate whether this speed justifies current investment levels. Some executives report that corporate boards are increasing their ambitions following early successes. Azhar also noted that even in slower markets like Italy, executives report rising trust and expanding budgets despite initial missteps.

Meanwhile, many organizations are moving away from expensive frontier models in favor of open-weight alternatives. In this scenario, AI usage might expand, but the financial bubble could still burst if insufficient revenue reaches the companies funding the underlying infrastructure. Azhar highlighted a Boston Consulting Group survey indicating that approximately 70 percent of CEOs worldwide believe AI success affects how others perceive their leadership. This dynamic creates an incentive for executives to present a rosier picture than reality supports.

Azhar acknowledged he lacks a simple answer regarding the future of the AI bubble and described the overall situation as finely balanced.