Impactive Capital合伙人公开质疑AI巨额资本开支回报，点名CoreWeave融资成本与Google自由现金流压力

It is "very hard to understand when a bubble might pop," said Lauren Taylor Wolfe, managing partner at Impactive Capital. Wolfe expressed deep uncertainty about the financial viability of massive artificial intelligence investments during an interview with CNBC. While acknowledging that AI has proven to be "an incredible tool" that "has enhanced productivity," Wolfe raised pointed questions about whether the technology can generate adequate investor returns. Wolfe emphasized the need to evaluate AI from multiple angles: consumer adoption, business models, and investment returns. "The bigger question is, does it have an effective business model, and will it generate investor returns?" she said. Goldman Sachs has estimated that hyperscalers will spend approximately $3T on AI capital expenditures between 2026 and 2028, but generating competitive returns on that investment would require between $5T and $10T in revenues. The managing partner drew a distinction between large hyperscalers and smaller "neocloud" providers, suggesting the latter face significantly higher risk. "With the neoclouds, they're borrowing; their cost of capital is going up, there's an inflating cost of compute," Wolfe explained. She noted that CoreWeave's ( CRWV ) borrowing costs have risen by 15%, while the prices customers are willing to pay for AI tokens continue to decline. Wolfe pointed to increasing scrutiny from corporate finance departments as a critical factor to watch. After treating 2025 and 2026 as "experimentation years," companies are now demanding accountability for AI spending. "The CFO office is putting a lot of scrutiny on budgets and how much return they're getting on their spend," she said, adding that if AI investments fail to generate revenues or reduce costs, "a lot of that is going to come to questions." When asked directly whether she believes AI investments will fail to generate returns, Wolfe acknowledged uncertainty but expressed growing skepticism. "I don't know the answer, and I think the jury's out," she said. "Given the quantum of capital, it's increasingly with the rising cost, it's increasingly less likely. There's very little margin of error, particularly for the neoclouds." Even major technology companies are showing signs of strain, according to Wolfe. She noted that this was "the first year, I think, that we saw Google ( GOOGL ), ( GOOG ) without free cash flow," suggesting that investors should apply "a microscope and a magnifying glass" to these developments. Artificial Intelligence/Robotics ETFs: ( AIQ ), ( BOTZ ), ( DTEC ), ( WTAI ), ( XAIX ), ( WISE ), ( GINN ), ( ROBT ), ( TECB ), ( XT ), ( THNQ ), and ( CHAT ).