Wall Street's forecasts for Big Tech require a major leap of faith: that the biggest AI hyperscalers can boost revenue much faster than the costs of running their businesses. Some of the numbers look too good to be true.
Investors are already on edge about these companies' soaring capital expenditures related to artificial intelligence. S&P this month cut Oracle's credit rating to one notch above junk. Alphabet shares slipped last week after the Google parent boosted its capex estimates. Microsoft and Meta Platforms report earnings Wednesday, while Amazon.com's results are due Thursday.
The bullish earnings narrative requires these companies to achieve newfound operating efficiency even as they spend trillions of dollars in capex over the next few years. The premise is evident in the consensus estimates for the five hyperscalers when combined as a group.
Analysts project that operating margins will improve to about 31% in calendar 2029 from 27% last year, according to data compiled by Visible Alpha. This would be the highest for the group collectively since Meta went public in 2012. Half of that improvement hinges on sales, general and administrative expenses declining to about 8% of revenue from 10% over the same period, while revenue almost doubles. The analysts' projections show each company's SG&A falling as a percentage of revenue, helping sustain or expand operating margins.
This scenario leaves little room for error. Depreciation expenses will soar in the next few years because of all the new buildings and equipment. Those are part of operating income under generally accepted accounting principles. And if those costs land higher than projected, relying on overhead to shrink as a share of revenue won't be enough to bridge the gap.
A two-percentage-point drop in SG&A might seem modest. But with combined annual revenue expected to top $3 trillion in three years, it compounds rapidly. The shift means the five hyperscalers would spend $77 billion less on corporate operations in 2029 than if overhead remained at its 2025 share of revenue. That efficiency squeeze is one of the crucial engines driving the margin forecasts.
But with some of the SG&A estimates, another possible explanation is that Wall Street analysts locked in their revenue and earnings estimates based on management guidance, then reverse-engineered the expense projections so they would fit.
The math behind the earnings can't work unless SG&A declines significantly as a share of revenue. So it is telling how analysts handle the expense lines in the middle.
Consider the analyst estimates for Oracle and how widely dispersed they are relative to the average. The standard deviation for their fiscal 2029 revenue forecasts is a narrow 3% of the average estimate, according to Visible Alpha. It is 4% for the nonstandard earnings metric that is tailored to match Oracle's long-term guidance.
But look at the divergence in the expense forecasts. The standard deviations swell to 17% for sales and marketing costs, 25% for general and administrative expenses, and 36% for depreciation. That indicates analysts are treating the middle of the income statement as a plug to reconcile the top and bottom.
All told, the consensus estimates envision Oracle's operating margins holding steady at about 31% each year through fiscal 2029 and improving to 33% in fiscal 2030. Revenue soars, driven by anticipated business with OpenAI, and so do cost of goods sold and depreciation. But SG&A, while climbing in dollar terms, stair-steps down to 6% of revenue in fiscal 2030 from 15% in fiscal 2026, which ended May 31.
Kevin Koharki, a Purdue University accounting professor and founder of CAE Consulting, says the SG&A estimates across the group strain credulity. "Whenever growth is this fast, there's got to be some strong positive correlation between SG&A and revenue," he said. "It just seems like the analysts, in order to keep the operating margins steady or slightly growing, are assuming that whatever depreciation increases by, we're going to see a direct offset through SG&A. I can't think of a scenario where that's ever happened before."
Implicit in the analysts' models is a belief that Big Tech can use AI to unlock rapid productivity gains, allowing revenue to grow much faster than overhead. Bringing in vast amounts of new business historically requires larger sales teams, customer support and administrative capacity. If AI tools fail to drive the hoped-for efficiencies, corporate expenses will scale with revenue, and the projected margin benefits won't materialize.
The expenses hit cash, not just earnings. Most of SG&A consists of cash outlays for costs such as payroll, commissions and marketing. Overruns would reduce the operating cash flow needed to fund the AI build-out, making it more likely the companies would have to issue additional debt or sell stock to cover shortfalls. Amazon and Oracle have already turned free-cash-flow negative. So has Alphabet, which halted stock buybacks and issued equity to fund its spending.
It will take more than spreadsheet mechanics to make these businesses truly more efficient.
Comments