Microsoft, Alphabet and Amazon report within 48 hours of a Federal Reserve decision, and investors are no longer asking if AI spending is large — they are asking whether big tech earnings week AI capex can finally pay for itself in cash.
Track it live: Live market tracker — indexes, megacap stocks and rates
This is the verdict week for the artificial intelligence build-out. Between October 26 and 30, 2026, the Federal Reserve will set interest rates, the government will publish third-quarter growth and the Fed's preferred inflation measure, and three of the world's most valuable companies will report earnings. Bloomberg, in reporting on October 26, framed the stakes directly: this week's prints will help judge whether more than $800 billion of expected hyperscaler capital expenditure can earn a return.
The phrase that matters — big tech earnings week AI capex — captures the collision. Earnings used to be graded mainly on revenue and profit per share. This week they will also be graded on free cash flow, debt, and what executives promise for 2027 spending. After Meta reported on October 21 and was punished for exactly those lines, no chief financial officer can pretend the question has gone away.
The calendar: Fed first, then Microsoft and Alphabet, then Amazon
The Federal Reserve announces its rate decision on Wednesday, October 28 at 2:00 p.m. ET. A few hours later, after the closing bell, Microsoft reports its fiscal first quarter of 2027 and Alphabet reports its third quarter of 2026. On Thursday, October 29, investors get third-quarter GDP and the September Personal Consumption Expenditures price index in the morning, then Amazon's third-quarter results after the close. Apple, which reports on Monday, November 2, sits just outside the week, while Meta and Tesla already reported on October 21.
Consensus compiled by SentiSense and read on October 25 sets a high bar. Microsoft is expected to post earnings per share of $4.71 on revenue of $90.63 billion. Alphabet is expected to post $3.03 per share on $127.91 billion. Amazon is expected to post $1.96 per share on $202.16 billion. Those are not distressed numbers. They describe three businesses still growing at a scale no previous generation of companies reached. The risk this week is not that the top lines collapse. It is that even strong top lines no longer settle the argument, because the spending underneath them has grown faster.
What Meta already told the market
Meta's print is the warning that hangs over the week. Revenue rose to $64.6 billion, up 26 percent year over year — by any historical standard, an excellent quarter. Yet quarterly capital expenditure reached $34.9 billion, free cash flow was negative $2.12 billion, described in market reports as Meta's first cash outflow in more than a decade, and management narrowed full-year 2026 capex guidance to $135 billion to $145 billion. The stock fell sharply, around 6 to 9 percent across reports, with individual reports ranging from 6.1 to 9.4 percent.
The contagion was immediate. On October 22 the Nasdaq fell 2.6 percent, its worst day since June, and the Philadelphia Semiconductor Index fell 6.7 percent. When one hyperscaler confirms AI infrastructure can consume more cash than the business generates in a quarter, investors reprice every other hyperscaler's promises the same morning. Microsoft, Alphabet and Amazon report in Meta's shadow.
Why this matters
Why this matters goes beyond one earnings season. The hyperscalers are now among the largest builders in the American economy. Their data centres consume power, land, chips, copper, cooling and construction labour at a pace that affects electricity prices, regional development and the federal borrowing environment. If the returns arrive — if cloud AI revenue grows fast enough to cover depreciation, power and financing costs — the build-out becomes the productivity story of the next decade. If returns lag, the costs do not disappear. They sit on balance sheets as depreciating servers, long leases and debt that must be refinanced regardless of model quality.
There is also a market-structure issue. The Magnificent Seven now trade below 24 times forward earnings, compared with a ten-year average near 29 times, according to Bloomberg figures cited via eToro, while expected 2027 earnings growth has been revised up to about 27 percent, from below 19 percent. The stocks have become cheaper relative to history just as analysts expect faster growth. Bulls call that a setup; bears call it a warning that the earnings depend on an AI cycle that has not yet proved it can fund itself.
Context and background: how the spending outran the cash
Two years ago, the large platforms could fund AI investment from operating cash flow and still buy back stock. That arithmetic has tightened, company by company. Cash-flow figures compiled from company filings via AIStockWire for the latest reported quarters make the divergence stark. Amazon generated $45.4 billion in operating cash flow and spent $54.2 billion in capex — spending exceeded operating cash by $8.8 billion. Alphabet generated $39.1 billion against $44.9 billion in capex. Meta generated $32.8 billion against $33.9 billion. Microsoft was the outlier in the right direction: $55.4 billion in operating cash flow against $35.8 billion spent, a cushion of nearly $20 billion.
Debt has filled part of the gap. In 2026, according to the same filing-based compilation, Amazon raised $67.0 billion in debt. Alphabet raised $56.2 billion in debt plus $49.6 billion in stock. Meta raised $24.9 billion in debt. Microsoft raised no new debt and kept its buybacks. That contrast is central to this week's grading. Microsoft can argue its AI programme is still self-funded. Amazon and Alphabet must argue that borrowing and, in Alphabet's case, issuing equity were opportunistic financing for assets that will throw off cash for years. Meta must argue its negative quarter was a timing trough, not a new normal.
Investors have noticed the shift in who benefits if growth disappoints. Last week small caps rotated higher, with the Russell 2000-tracking IWM up 2.0 percent while the Nasdaq-100-tracking QQQ fell 0.8 percent. A 2.8-point gap in one week does not end megacap dominance, but it shows capital looking for earnings that do not require another data centre to materialise. The Fed decision sharpens that choice: higher-for-longer rates raise the discount rate on distant AI profits and raise the interest bill on the new debt itself.
The bull case: demand is real and capacity is sold
The bull case is not imaginary. Supporters point to Azure growth targets around 45 percent in constant currency for Microsoft, to Amazon chief executive Andy Jassy's statements that AWS capacity is sold out through 2027, and to Google Cloud's 35.6 percent operating margin, which challenges the claim that cloud AI is inherently unprofitable. If enterprises are signing multi-year contracts, if inference demand keeps compounding, and if each new generation of accelerators lowers cost per token, then today's capex is tomorrow's high-margin revenue — the same pattern that once made investors forgive Amazon for warehouses and Google for its early data centres.
Bulls also argue the valuation already reflects doubt. Below 24 times forward earnings, against a ten-year average near 29 times, is not bubble pricing for companies expected to grow earnings by roughly 27 percent in 2027.
The bear case: free cash flow is the report card now
The bear case starts with Meta's negative $2.12 billion free cash flow and asks how many such quarters the group can post before buybacks shrink and equity holders demand a pause. Some analysts cited in recent market coverage have forecast negative group free cash flow by December 2026 if current spending trajectories hold. Even if that proves too severe, its direction explains the new grading system: beating earnings per share by a few cents will not rescue a stock whose management raises 2027 capex again without naming the contracted revenue behind it.
Critics also question depreciation. Servers may lose economic value faster than accounting schedules imply if a new chip generation cuts inference costs sharply. A data centre shell lasts decades; the silicon inside it may not.
What happens next
What happens next unfolds in three acts. First, the Fed at 2:00 p.m. ET on Wednesday sets the financing backdrop: a hawkish hold or hike talk makes every 2027 capex promise more expensive. Second, Microsoft and Alphabet on Wednesday evening, followed by Amazon on Thursday, must answer three questions in order — did cloud revenue accelerate, did free cash flow hold, and is 2027 spending guided to revenue that already exists? Third, Thursday's GDP and September PCE data will tell investors whether the wider economy can keep absorbing higher power, construction and borrowing costs generated by the build-out.
The most likely outcome is differentiation, not a uniform verdict. Microsoft's self-funded position gives it the widest margin for error. Alphabet must reconcile strong cloud margins with its debt and equity raising. Amazon must show that spending $54.2 billion against $45.4 billion of operating cash flow was the peak of the strain, not its pattern. Apple, reporting November 2, will then face whether staying lighter on capex was discipline or delay.
For readers following big tech earnings week AI capex as an economic story, the number to watch is not any single earnings-per-share beat. It is the gap, company by company, between operating cash in and capital out. When that gap closes because revenue caught up, the build-out will have passed its hardest test. Until then, every earnings week is also a financing week.
Related coverage: Meta's record revenue and negative free cash flow, explained and Apollo's warning that AI gains have bypassed most of the S&P 500.