Wall Street Celebrates Big Tech's AI Growth, But Concerns Bubble Under The Surface

Wall Street’s tech “Super Bowl” Wednesday evening was widely considered a big win for AI bulls, with Microsoft, Google, Amazon and Meta all reporting higher-than-expected revenue growth tied to their artificial intelligence investments.

But while some industry observers framed tech’s strong quarter as vindication that the data center spending spree is paying off, others are pointing to red flags buried beneath the top-line numbers and warning that it is too soon to take a victory lap. 

Google, Amazon, Microsoft and Meta made a combined $130B in capital expenditures in the first three months of 2026, a 71% increase from the year prior. And their quarterly earnings suggest this unprecedented spending on AI infrastructure is paying dividends. 

The three largest cloud providers all surpassed analyst estimates last quarter, buoyed by AI-linked gains. Google Cloud’s revenue jumped 63% over the prior year, while revenue from its generative AI products grew 800%. Microsoft’s cloud business grew faster than expected, at 40% year-over-year, while AWS’ 28% growth was more than a billion dollars above Wall Street estimates. 

The overall cloud market grew by 35% year-over-year, according to Synergy Research Group, the industry’s highest quarterly growth rate since 2021, when it was just 40% of its current size. Meta, which doesn't provide cloud services, also exceeded revenue expectations amid aggressive AI spending, with 33% growth.

Following these results, the four Big Tech firms plan to ramp up AI capex even further this year to more than $725B, a more than 60% increase from 2025. 

The first-quarter tech earnings have largely been treated as a triumphant moment for a sector that in recent quarters has been mired in a narrative of an AI bubble on the verge of popping.

“Wow, that was some quarter,” Synergy Chief Analyst John Dinsdale wrote in an email. “Take a bow, AI. Ever since ChatGPT was launched in late 2022, the cloud market has gone into overdrive.

“Our forecasts point to sustained strong growth in the years ahead, with AI continuing to drive usage, unlock new use cases, and boost cloud provider revenues.” 

Yet other industry analysts caution that these striking top-line numbers shouldn't blind investors — or the tech sector itself — to potential headwinds that were also visible in this week’s earnings. They suggest the AI boom may rest on shakier ground than many AI boosters acknowledge. 

One trend generating concern is that while hyperscalers are spending more than ever on AI infrastructure, they are getting progressively less bang for their buck. 

The cost of building and leasing new data centers is going up, the result of supply constraints putting upward pressure on prices for everything from electricity and critical equipment to optical fiber and labor. Prolonged project timelines due to rising community opposition and legal hurdles have also contributed to higher costs.

The AI computing equipment housed inside data centers is also getting more expensive. Shortages of graphics processing units and other chips and information technology gear needed for AI are rising rapidly amid unyielding demand.

These dynamics have had a disparate impact across the digital infrastructure ecosystem.

Equipment suppliers and data center providers have been able to pass higher costs along to end users and have benefited from the supply-demand imbalance. But for the hyperscalers themselves, rising costs simply mean less compute for each AI dollar they are being asked to justify. 

“We’re seeing constraints across the board. The hyperscalers who are trying to get into the gold mine — they’re having to wait, or spend more to get in,” Jefferies analyst Brent Thill told The Wall Street Journal. “It’s good for the picks and shovels, but it’s not good for the people who are assembling all the pieces.”

The four major tech firms aren't impacted equally by these cost challenges. In fact, some have turned them to their advantage.

Google and Amazon produce their own chips and have begun selling them to other major AI infrastructure customers. In doing so, they have turned a source of rising costs into new revenue. 

But Meta and Microsoft are chip customers and are exposed to these cost pressures without benefiting from them. This difference, analysts said, is at least partially responsible for the divergence in Big Tech stocks. 

Shares in Google parent company Alphabet rose 10% after reporting earnings. Amazon rose 0.8% Thursday and was up 27% from the start of April. Meta's stock price fell 9% Thursday, and Microsoft dropped 4%, but both were still up from the beginning of the month. 

“If you’re a cloud services company and you have a full suite of computing services, from the chip down to the model, down to the application, you’re doing much better, versus if you’re just building data centers and running more third-party models,” John Belton, a portfolio manager with the Gabelli Growth Fund, told The Wall Street Journal. 

Hyperscalers are also facing concerns about the share of these record cloud revenues that are tied to a small number of customers, particularly AI-model makers OpenAI and Anthropic.

To some analysts, this represents a dangerous concentration of risk.

Microsoft has generated the most anxiety around its reliance on a single customer. It acknowledged this week that nearly 45% of its cloud revenue over the past two quarters has come from OpenAI, with whom it maintains an infrastructure partnership and ownership stake. While Microsoft’s ties to the ChatGPT maker boosted its fortunes in the early days of the AI race, analysts are increasingly pointing to it as a liability. 

“MSFT's investment in OpenAI was initially considered a masterstroke of corporate strategy,” Sarfatti Investment Research wrote in a note last week. “However, I believe this partnership has turned into a significant risk of counterparty concentration.”

Although Microsoft addressed some of these concerns through a restructuring of its OpenAI partnership last month, the other major cloud providers face similar questions.

The fortunes of Oracle, which is set to report earnings March 10, are similarly tied to OpenAI through the Stargate cloud and data center development partnership.

While Amazon hasn't disclosed the exact breakdown of its cloud demand, the company recently signed hundred-billion-dollar deals with OpenAI and Anthropic, and the company’s leadership was pressed by analysts Wednesday for insight into how much of its cloud business is reliant on AI-model makers versus a wide array of corporate customers. 

Questions around risk concentration have become more pronounced as both major AI-model makers have run into turbulence.

Anthropic has been blacklisted by the Pentagon and is now engaged in a legal battle with the Trump administration after a dispute over putting guardrails on the military's use of its AI models. 

OpenAI is facing questions about its future even as it ramps up to an expected initial public offering. Last month, reports emerged that the firm missed internal revenue targets and saw growth slow dramatically in late 2025. And OpenAI has backed out of several data center deals in recent weeks, sparking speculation that it is cutting back on its planned infrastructure build-out. 

These issues have triggered Wall Street fears that any significant pullback from just one of these companies could pull the rug out from under cloud providers and the AI sector as a whole.

“On the surface, yes it’s a concern, as any tenant concentration poses risk,” David Guarino, who leads Green Street’s data center research, said in an email to Bisnow. “Any failure/struggle of one of those two companies would have very bad implications for AI as a whole, which no hyperscaler wants.”

Still, Guarino said he isn't particularly concerned that this kind of chain reaction will unfold. He remains bullish on OpenAI and Anthropic, two “enviable” companies with “very deep pockets.” 

Additionally, the major cloud providers are aware their futures are wedded to these companies, and some have direct ownership stakes in them, Guarino said.  These investment-grade hyperscalers with lots of cash on their balance sheets are likely willing to do whatever it takes to ensure OpenAI and Anthropic are able to survive any slowdown in the near future. 

“There are many parties who are highly incentivized to see Anthropic and OpenAI succeed,” Guarino said. “When you dig past the surface, the concentration risk isn’t as concerning as it first appears.”