Elkann: AI Buildout Needs Revenue Models or a 10-Year Mismatch Looms

For AI engineers and product leaders, Elkann's 'supply and demand mismatch' is not just a market risk — it's a product-market-fit problem at the scale of the entire AI buildout. When a Meta board member and industrial chairman says revenue-generating mechanisms are 'less certain,' the signal is that AI infrastructure may be running ahead of applications that can pay for it.

Ferrari chairman John Elkann used an exclusive interview with CNBC at the Wave 2026 technology conference in Turin on Friday, October 9, 2026, to deliver a dual message on artificial intelligence: he is not in the bubble camp, yet the mismatch between AI's expanding supply and the demand side's monetization models is where market risk could concentrate. Elkann, who chairs Ferrari and Stellantis, sits on Meta Platforms' board, and invests through family holding Exor, occupies a rare seat at the intersection of industrial manufacturing, digital platforms, and family-office capital. That vantage point gives his comments more weight than a typical executive airing tech optimism.

“

Elkann, who chairs Ferrari and Stellantis, sits on Meta Platforms' board, and invests through family holding Exor, occupies a rare seat at the intersection of industrial manufacturing, digital platforms, and family-office capital.

His core statement — "I am not in the camp of the bubble believers" — was coupled with a warning: "What is less certain is what are the revenue-generating mechanisms and business models out of that technology. That's where you need to see where the mismatch can happen." He elaborated that bubbles come in different forms, distinguishing between those driven purely by financial speculation and those fueled by rapid buildup of investment and infrastructure. His emphasis on the technology's "direction of travel" suggests he views AI's capital cycle as structural rather than anomalous, even as near-term multiples may be tested.

The comments land amid intensifying investor debate over AI capex. In 2025 and 2026, hyperscalers and chipmakers have poured hundreds of billions into data centers, GPUs, energy, and networking, yet recurring enterprise revenue has not kept pace in every segment. Elkann's observation about high energy and compute requirements adds an operational lens: the supply side is real and expensive, but the demand side—the applications and business models that justify that spending—remains less certain. This is not a call for an imminent crash; it is a warning about the potential for repricing if enterprise adoption and pricing power fail to catch up to infrastructure capacity.