Wall Street’s latest Big Tech sell-off reflects growing unease over the staggering sums of cash companies are burning to fund an artificial intelligence infrastructure build-out that has yet to yield commensurate returns.
At the same time, anxiety is growing that even the world's largest tech companies may not have pockets deep enough to build the data centers and other infrastructure they say will be needed to meet future demand.
While rising public opposition has grabbed national attention, some analysts suggest the biggest issue that may ultimately slow down the data center boom is tech companies’ ability to raise capital.
)
Quarterly earnings from Meta, Alphabet, Microsoft and Amazon triggered one of Wall Street’s most dramatic sell-offs of the AI era last week. By the time markets closed Friday, the four companies that have driven much of the ongoing AI data center boom had collectively lost close to $900B in market capitalization, and shares of all four firms remained below pre-earnings levels nearly a week later.
Google parent Alphabet suffered the steepest decline among the hyperscalers, its stock falling more than 13% following its earnings release. The drop reduced the firm’s market capitalization by close to half a trillion dollars, marking one of the worst stretches for the stock since the company went public.
Meta has also fallen sharply, with shares down close to 15% over the past month. Amazon stock fell more than 5% in the days following earnings, while Microsoft — the strongest performer of the group — still posted a roughly 3% decline.
“The tech pool is closed,” Jefferies analyst Brent Thill told the Financial Times. “We’re seeing a broad sell-off with everyone running right now. ... We don’t know if the closed sign will be up for 10 minutes, 20 minutes or longer.”
At the heart of Wall Street's reaction is AI infrastructure spending that continues to accelerate, even as revenue growth tied to that investment slows and tech giants burn through cash at an alarming pace.
Meta, Alphabet, Microsoft and Amazon collectively spent close to $207B on data centers, chips and other capital expenditures last quarter, nearly double the $113B they spent a year earlier. That figure is expected to rise again next quarter, as hyperscalers plan to spend a combined $232B, which would push annual capex close to $740B.
But AI-related capex has been climbing for the better part of four years. What has changed is that the most cash-flush companies in the world are burning through their coffers to fund it.
Until recently, the major hyperscalers largely funded their infrastructure build-outs from operating cash flow, with overall earnings still outpacing the escalating AI price tag.
Last quarter, however, Alphabet, Meta and Amazon together pushed into negative territory, spending $45.2B more than they generated in free cash flow. Just a year earlier, these same firms reported positive cash flow to the tune of $36.4B.
Alphabet emerged as the starkest example of investor concerns over negative cash flow tied to slowing AI returns. Its $51.4B in quarterly capex pushed its free cash flow to negative $9.1B, down from positive $24.5B a year prior. At the same time, revenue growth from its critical search business slowed for a second consecutive quarter as the firm struggled to monetize AI integration, fueling concerns that AI spending hasn’t translated into commensurate revenue gains. These worries overshadowed Google’s cloud revenue nearly doubling year-over-year.
“Investors are increasingly unhappy at the huge sums of money being spent on AI infrastructure without a clear road map for when, or whether, those investments will deliver a sufficient return,” said Dan Coatsworth, head of markets at AJ Bell, according to a Morningstar investor memo. “That means investors are increasingly questioning if money is being spent wisely.”
Google was far from alone. Meta similarly posted negative free cash flow of $2.1B after $34.9B in quarterly capex. Amazon spent the most of the group, with a record $75.8B in quarterly capex pushing cash flow to negative $33.5B. Amazon Web Services sales jumped 39%, but that growth was slower than a year prior and failed to calm investors.
Even Microsoft, which remained cash flow-positive as cloud revenue growth increased 31%, was pulled lower by investor concerns around the health of the AI segment as a whole.
“Last week’s earnings highlighted a growing tension between strong demand for AI-linked goods and services and the rising cost of meeting it. … Each additional dollar of revenue is generating less additional profit, raising questions about whether returns can keep pace with soaring investment,” BlackRock analysts wrote in a note to investors Monday. “But whether that demand will be enough to justify the scale of investment is still an open question.”
)
Leaders of the major hyperscalers have responded to these investor concerns much as they have throughout the AI boom, urging Wall Street to trust that escalating infrastructure spending will ultimately pay off.
Alphabet CEO Sundar Pichai and Amazon CEO Andy Jassy framed the moment as a once-in-a-lifetime technological shift in which there is no choice but to invest aggressively now to build the foundations of an AI future.
Despite Amazon’s negative cash flow, Jassy told analysts last week that the firm’s leaders “have even more conviction than we already had that the investment we're making in AWS will yield a much larger revenue and free cash flow-generating business than we would have imagined a couple of years ago with strong return on invested capital.”
“This is a once-in-a-lifetime opportunity — what we're seeing with AI, it's the biggest transformation in our lifetime,” Jassy said. “Would it be more convenient if this type of acceleration and opportunity happened over 25 years and you could spread the capex over 25 years? Yes, it might be more convenient, but that's not the world that we live in, and that's not the opportunity.”
Yet Big Tech’s path toward its AI future faces headwinds beyond just stock market skepticism of the thesis underpinning its infrastructure spending.
There are growing questions about the sector’s ability to continue funding its infrastructure build-out long enough for promised returns to materialize.
Negative cash flows mean tech giants are increasingly looking for capital infusions from the debt and equity markets to help fund their build-out, but this need is emerging just as financing AI infrastructure is becoming more difficult — and more expensive.
Just as equities traders are demanding stronger evidence that rising capex will translate into earnings, lenders are becoming more selective about financing new infrastructure. This shift has been most apparent in structured debt markets, where yields on bonds tied to AI infrastructure have skyrocketed in recent weeks.
This has been particularly pronounced for debt tied to neoclouds, AI labs and other higher-risk firms. CoreWeave bonds issued earlier this year at a 9.75% rate now trade at a 15% yield. The massive cost of capital means the AI cloud provider has effectively been shut out of the bond market, according to The Wall Street Journal, as have other neoclouds like Nebius and Nscale.
“The bar is going higher and higher in terms of proving to lenders that this investment will turn into sufficient revenue,” Pimco Chief Investment Officer Dan Ivascyn said, according to the WSJ. “The weak links are going to suffer first.”
While investment-grade giants like Amazon, Alphabet and Microsoft had been insulated from the credit market volatility hitting AI startups, yields on bonds tied to major hyperscalers also began to climb last week, diverging from the investment-grade indexes they typically track closely.
Analysts said there are several reasons debt markets are no longer as forthcoming with liquidity. In part, it is the same concern driving these firms’ share prices down: growing questions over how soon infrastructure investments will turn into sufficient revenue.
But analysts said even true believers in tech firms’ AI strategies are concerned over concentration risk. The scale of what hyperscalers are asking for is unprecedented, and there is only so much capital that any lender can have tied up in just a handful of companies all pursuing roughly the same business model.
With wallets tightening, analysts said hyperscalers will be forced to get increasingly creative in the coming months to access the capital to fund their AI build-outs.
Alphabet and Meta are leading the charge on this front as their cash flows turn negative, with both issuing new equity. Alphabet is issuing new shares to raise close to $85B, its first share sale in 20 years. Meta announced last week it is trying to raise $95B by issuing new nonvoting shares.
“Everybody is trying to raise money for their build-outs before the liquidity dries up,” Carnegie Investment Counsel Chief Investment Officer Ben Connard said, according to The Wall Street Journal.
“Ultimately, there might not be enough money to fund this build-out,” he said. “Financing could be the ultimate crunch that slows things down.”
)
)
)
)
)
)
)
)
)
)
)
)