If you've been watching the AI trade lately, you've seen a lot of green. On Tuesday, eight AI-focused exchange-traded funds hit fresh 52-week highs, riding a wave of investor enthusiasm that's spreading well beyond the usual suspects like Nvidia (NVDA). The list includes funds like the KraneShares Artificial Intelligence & Technology ETF (AGIX), the Global X Artificial Intelligence & Technology ETF (AIQ), and the Amplify Bloomberg AI Value Chain ETF (AIVC), among others.
But here's the thing: while markets are cheering, corporate America is quietly sweating. Companies are spending aggressively on AI tools, tracking everything down to the cost of individual prompts—those little "tokens" that power chatbots and automation. Yet most of them can't actually prove that all this spending is making them any more money.
A 2026 survey by ModelOp, an AI lifecycle management and governance platform, found that more than two-thirds of companies rely on squishy metrics like "time saved" rather than hard earnings impact to justify their AI investments. ModelOp calls this the "AI value illusion," and it's a pretty apt name. It's like buying a fancy espresso machine for the office and claiming it boosts productivity because people seem perkier—without actually checking if sales went up.
So far, the Magnificent Seven have been the exception. Microsoft (MSFT) and Alphabet (GOOGL) have managed to show real results from their AI spending. Microsoft delivered triple-digit AI revenue growth in its latest quarter, powered by solid cloud performance. Alphabet's massive capital expenditure push is backed by what Direxion's senior executive Ryan Lee calls a "$460 billion order backlog," alongside 22% revenue growth. That's the kind of math that makes investors feel good.
But for most companies, the story is murkier. Even as tools from Microsoft and others let firms track usage metrics like prompt volume and active users, executives admit that attribution—proving AI caused a revenue bump—is the hardest problem. AI might correlate with productivity gains, but isolating it as the driver is like trying to figure out which ingredient made the soup taste better.
This disconnect matters for ETF investors. Funds like the Vistashares Artificial Intelligence Supercycle ETF (AIS) and the FT Bloomberg Artificial Intelligence ETF (FAI) are increasingly tilted toward "AI adopters"—companies embedding AI into their workflows—rather than just infrastructure providers like Nvidia. But if adoption can't be tied to revenue or productivity, the next leg of the AI trade could face some serious scrutiny.
The rally suggests markets are pricing in a future where AI-driven productivity becomes measurable and material. Yet current data tells a more cautious story. According to McKinsey & Company, cited by CNBC, while 64% of companies say AI is driving innovation, only 39% report a tangible earnings impact. That's a gap big enough to drive a truck through.
That tension could define the next phase of AI ETFs. For now, investors are rotating into the full value chain via funds like the iShares Robotics and Artificial Intelligence ETF (ARTY) and the Tcw Artificial Intelligence ETF (AIFD). But until corporate ROI catches up with adoption, the AI trade may remain as much about expectation as execution. And as any seasoned investor knows, expectations have a way of resetting when the numbers don't show up.













