Big Tech Earnings Week — $725 Billion in AI Spending Has to Prove Itself. Right Now.

For the past three years, the largest technology companies in the world have had an unspoken agreement with Wall Street: spend whatever it takes on AI infrastructure, and the market will reward you for it. That agreement is breaking down. For years, US technology giants had a tacit agreement with investors — the companies could spend lavishly on artificial intelligence, and the stock market would reward them as long as their revenues were rising. That deal is suddenly breaking down.

The proof arrived last week. Google posted its strongest quarter in years — Google Cloud revenue up 82% year over year to $24.8 billion — and the stock still fell more than 6% the morning after results dropped. The reason was not the revenue. It was the number that came with it: capital expenditure of $44.9 billion in a single quarter, a figure so large it pushed Google's free cash flow negative for the first time in its 22-year history as a public company. Investors had expected a beat. They got one. They sold anyway.

Today, July 29, Microsoft and Meta report after market close. Apple and Amazon follow tomorrow. Together, these four companies are on track to spend a combined $725 billion on capital expenditure in 2026 — a 77% increase from the $410 billion they deployed in 2025. Every one of those dollars needs to demonstrate a return, and the market has made clear it is no longer willing to wait indefinitely for the proof.

What Changed — Why a Revenue Beat Is No Longer Enough

The Alphabet moment last week was a turning point. Five days ago, Alphabet raised its 2026 capital expenditure guidance to as much as $205 billion, and its free cash flow turned negative for the first time in the company's history as a public company. Investors had expected a cloud beat. They got one — Google Cloud revenue surged 82% year over year to $24.8 billion — and they still sent the stock down more than 6% the following morning.

The message from that sell-off is the context for everything that reports this week. Investors are no longer pricing these companies on revenue growth alone. They are pricing them on the relationship between revenue growth and the capital being consumed to generate it. When a company spends $44.9 billion in a quarter and generates negative free cash flow, the question becomes: when does this spending pay back? If the answer is "in a few years," the market's next question is: at what interest rate, and what is the probability the AI revenue curve holds that long?

That repricing of the AI narrative is visible in the stock performances heading into this week. Despite strong AI and cloud demand, MSFT and META stocks are down 20.7% and 9.7%, respectively. These are not companies that have missed expectations. They are companies that have met or beaten them — and still seen their valuations compressed because the market has decided the spending trajectory is the story, not the revenue line.

Microsoft — Azure Growth Is the Only Number That Matters

Microsoft reports fiscal Q4 2026 results tonight. Analysts are looking for earnings per share between $4.22 and $4.24, with revenue projected in the range of $87.5 billion to $87.67 billion. The EPS and revenue figures will receive a paragraph each in tomorrow's coverage. Azure cloud growth will receive everything else.

Management guided for Azure to grow 39 to 40% in constant currency for the quarter. The market has established a clear floor: any print below 36% risks triggering a sell-off regardless of headline EPS. The logic is straightforward — Microsoft has committed to approximately $190 billion in capital expenditure for FY2026, and much of that spending is AI infrastructure that serves the Azure customer base. If Azure growth decelerates, the return-on-investment case for that spending becomes structurally harder to defend.

The secondary story is Copilot. Microsoft's AI productivity suite now has more than 20 million paid seats — a number that represents real commercial adoption, not just enterprise trials. Whether those seats translate into measurable margin contribution will colour the guidance commentary. Microsoft's forward P/E has compressed to approximately 20.6x — near a multi-year low — which means significant bad news is already priced in, but also that the upside reaction to a genuine Azure beat could be significant.

The OpenAI dimension hangs over the print as well. Azure is the commercial backbone of OpenAI's products, meaning Microsoft's AI revenue is substantially exposed to OpenAI's customer growth. If OpenAI's Ohio campus deal with Nvidia and SoftBank eventually moves compute off Azure, that revenue leakage will appear in Azure growth numbers first.

Meta — $125 to $145 Billion in Capex and One New Chip

Meta also reports tonight, and its story is simpler to frame but harder to evaluate. Wall Street expects Meta to report EPS of $7.18 for Q2 2026, implying slight growth from the prior-year quarter. Revenue is estimated to rise about 27% to $60.22 billion. Meta's advertising business is growing strongly — ad impressions up 19% and ad pricing up 12% — which means the top line is not the concern.

The concern is what Meta is spending. Earlier this year, Meta raised its 2026 capital spending forecast to $125 billion to $145 billion, up from $115 billion to $135 billion. This is a company that generates most of its revenue from advertising, ploughing a sum larger than the entire GDP of Hungary into AI infrastructure in a single year. The justification is competitive survival — Meta's argument is that it must own the compute to train and serve the Llama model family, power Meta AI across WhatsApp, Instagram and Facebook, and build toward what Zuckerberg calls the long-term metaverse infrastructure. Whether the market accepts that argument tonight will depend heavily on whether operating margins hold above Q1's 41.4%.

The new hardware story at Meta is its in-house AI chip. Investors will also watch for updates on Meta's AI infrastructure strategy, including the rollout of its in-house "Iris" AI chip, which is expected to enter production in September. The chip is part of Meta's plan to expand its AI computing capacity to 14 gigawatts by 2027. An Iris update that shows the chip on track for September production reduces Meta's Nvidia dependency and strengthens the capex efficiency case — which is exactly what sceptical investors need to hear.

Apple and Amazon — Different Tests, Same Underlying Question

Apple reports tomorrow and faces a question distinct from its peers: it has spent less on AI infrastructure than any of the other four, and has instead bet on on-device AI — Apple Intelligence — as its differentiation strategy. The thesis is that consumers will upgrade their iPhones specifically for AI features, driving a hardware refresh cycle that lifts revenue without requiring the kind of cloud capex its rivals are committing to. Apple needs to prove that AI can drive an upgrade cycle for devices, rather than replicating its cloud services model. The expected EPS is $1.89, up 20.4% from the prior year. Whether Apple Intelligence is actually driving upgrade decisions — or whether it remains a feature people have heard of but not acted on — is the question that will define the stock's trajectory for the rest of 2026.

Amazon's story is the most operationally complex. AWS grew revenue 28% year over year last quarter, maintaining the kind of cloud growth rate its rivals would accept gratefully. The problem is what that growth cost: Amazon's free cash flow dropped to $1.2 billion last quarter due to investments. For a company of Amazon's scale, $1.2 billion in free cash flow is effectively zero. The AI infrastructure buildout is consuming capital at a rate that makes the business look far leaner on cash than its revenue line suggests. AWS operating income is what investors will study most carefully — if margins are compressing under capex pressure, the bull case for Amazon's AI position weakens significantly.

The Fed in the Background — Rates Hold but the Signal Matters

One more variable runs underneath all of this: the Federal Reserve. According to the Federal Reserve's official FOMC calendar, a meeting is taking place from July 28 to 29. The market broadly expects the Fed to hold interest rates steady, with a target rate between 3.50% and 3.75%. A hold is expected and already priced. What is not fully priced is the Fed's commentary on the inflation and growth outlook. If the statement sounds hawkish — emphasising upside inflation risks — the reaction in rate-sensitive tech stocks could amplify whatever earnings move was already coming from Microsoft and Meta tonight. Higher rates increase the discount applied to future cash flows, which directly hurts companies whose AI spending paybacks are years away.