A record share of fund managers now believe companies are overinvesting in artificial intelligence, a sentiment fueled by a recent $700 billion market dip triggered by a speculative report published on Substack, according to reports from Fortune and the Wall Street Journal.
The concerns center on massive, undisclosed financial commitments made by the largest technology companies – Amazon, Meta, Alphabet, Microsoft, and Oracle – to build out the data center infrastructure necessary to power AI development. Moody’s Ratings recently reported these five hyperscalers have accumulated $662 billion in future data center lease commitments not yet reflected on their balance sheets, representing 113% of their adjusted debt, as detailed in a February 25th report.
This accounting quirk, where future obligations aren’t immediately recognized as liabilities, has raised concerns among analysts. David Gonzales of Moody’s told Fortune the figure isn’t an avoided liability, but rather “yet to be on the balance sheet,” as the services haven’t been received to trigger the obligation. The total undiscounted future lease commitments for these companies reached $969 billion as of the complete of 2025.
The scale of investment is unprecedented. Spending by Google, Meta, Amazon, and Microsoft is projected to reach at least $670 billion this year alone, exceeding the financial commitment to the expansion of the U.S. Railroad system in the 1850s or the Apollo program in the 1960s.
Amazon, in particular, is facing scrutiny regarding its ability to sustain growth in its Amazon Web Services (AWS) division while defending margins in its retail business, according to a February 11th report from Forbes. The company’s operating margin of 11.2% lags significantly behind Microsoft’s software-driven 46.7%.
The surge in investment is similarly impacting the software sector. Components of the State Street software ETF have lost a combined $1.6 trillion in market capitalization this year, as investors question the long-term viability of traditional software in the face of AI disruption.
The influence of these tech giants extends beyond financial commitments. They spent more than $100 million in 2025 influencing government policy, the first time exceeding that threshold, potentially slowing down AI regulation.
Internally, Amazon is grappling with the impact of AI on its long-held culture of deliberate, thoughtful writing. Leadership is now encouraging employees to utilize AI for writing tasks, a shift that worries some who believe writing is integral to the company’s decision-making process. “Writing is thinking,” a longtime Amazon veteran told Kristi Coulter, who reported on the internal concerns. “I can’t tell you how many times I changed my mind when writing a narrative.”
Beyond the U.S., India is positioning itself as a key player in the data infrastructure landscape. The country produces 20% of global data and has a 62% generative AI adoption rate, but currently hosts only 3% of global data. Prime Minister Narendra Modi has articulated a vision for India to become a central hub for data hosting and servicing, guided by the principle of “Welfare for All, Happiness of All.”
A recent survey revealed that 83% of senior enterprise infrastructure executives believe their systems will fail within two years without major data infrastructure upgrades, with a third anticipating failure within the next year. This underscores the critical importance of data – and the infrastructure supporting it – as the foundation for AI adoption.
