The $335 Billion Question: Microsoft and Meta Both Report Earnings July 29 in AI's Biggest Test

Two of the world’s most important companies in the artificial intelligence buildout will report quarterly earnings on the same day, July 29—and the results will likely define how the market views the AI spending thesis for the remainder of 2026.

Microsoft, reporting its fiscal Q4 (ending June 30), and Meta, reporting calendar Q2 (ending June 30), together represent more than $335 billion in planned 2026 capital expenditures—a combined war chest being deployed into AI infrastructure, chips, and model development at a pace the technology industry has never seen. What they say about returns, demand, and outlook will matter more than almost anything else this earnings cycle.

Microsoft: The $37 Billion Run Rate Under Pressure

Microsoft enters the report with a complicated narrative. Its AI business crossed a $37 billion annual revenue run rate in Q3 FY2026 (ended March 31), up 123% year-over-year—an extraordinary number by any prior standard. Azure cloud, the primary vehicle for AI monetization, grew 40% in Q3. Guidance for Q4 calls for Azure growth in the range of 39% to 40% at constant currency, with revenue consensus at $87.7 billion (15% growth) and EPS of $4.24.

But the stock is down roughly 18% year-to-date. The market’s concern is not the numbers themselves—which have consistently beaten estimates—but the cost of producing them. Microsoft expects to spend approximately $190 billion on capital investments across all of 2026, including GPU clusters, data centers, and network infrastructure to power Azure AI services. Copilot has crossed 20 million paid seats (up 250% year-over-year), but seat count growth has slowed sequentially, raising questions about whether the core product is approaching saturation or simply transitioning to deeper enterprise integration.

The key questions analysts will be watching: Does Azure sustain 39-40% growth given the capacity Alphabet and others are adding as competition? Do Copilot usage metrics suggest deepening enterprise integration, or mostly experimental deployments that haven’t committed to long-term contracts? And—the question that will not be explicitly asked but will be read in every forward indicator—when does the $190 billion capex begin showing up as accelerating cloud margin?

One wildcard: Microsoft’s MAI family of proprietary models has been quietly displacing OpenAI and Anthropic in Excel and Outlook integrations, reducing its royalty obligations to external AI providers. If MAI-driven product metrics show strength in Q4, it could signal that Microsoft is building a defensible in-house AI stack rather than permanently depending on expensive external model licenses—a shift with significant long-term margin implications.

Meta: The $145 Billion Bet on a Future Platform

Meta’s situation has a different texture. The company raised its 2026 capex guidance to $125-145 billion earlier this year, citing accelerated infrastructure construction for Llama 5 training and its agents platform. Q2 consensus calls for revenue of approximately $60.17 billion (+26.6% year-over-year) with EPS of $7.13 per share.

The macro tailwind is real. Advertising markets have remained resilient in H1 2026, and Meta’s AI-powered ad targeting—built on its Andromeda recommendation system—has continued to deliver superior returns-on-ad-spend for performance marketers. This drives the ad price inflation that underlies Meta’s revenue growth even as user growth plateaus. The integration of AI features into WhatsApp, Instagram, and Messenger continues to expand engagement metrics, keeping users on platform longer.

The headwind is that Meta’s core advertising business is financing a $145 billion capital bet on AI infrastructure that has, so far, not produced a discrete, auditable revenue line in the company’s financials. Meta AI (the consumer chatbot), Llama model licensing, and the newly launched Meta Muse image and video generation platform are all real products—but none has yet contributed a material share of the company’s $60 billion revenue base. When Zuckerberg says Meta is building the platform for the agent economy, investors are being asked to trust a multi-year roadmap backed primarily by a social media advertising machine.

The critical question for Meta’s Q2 report is not whether revenue beats the $60 billion consensus. It likely will. The critical question is whether Zuckerberg’s commentary on agents—their timeline, their monetization mechanism, their enterprise traction—provides enough specificity to validate the capex level or triggers another round of “show me the revenue” skepticism from institutional holders.

The Context: A Week of Reckoning

The timing of the July 29 dual report day is not coincidental—it reflects standard fiscal calendar rhythms—but it arrives at a moment of heightened scrutiny about big tech AI spending.

The week’s context:

• Alphabet (July 22): Revenue beat, Google Cloud +82%, but first-ever negative quarterly free cash flow at -$5.9 billion on $44.9 billion in capex. Stock fell 4% after-hours.
• Intel (July 25): Stronger-than-expected Q2 on AI infrastructure demand, guidance raised. A positive data point suggesting chip orders are real and growing.
• Tesla (July 24): Q2 beat on robotaxi deployment and Optimus production progress, validating the AI-adjacent hardware thesis.

Microsoft and Meta will determine whether the week ends with a narrative of AI investment validated by accelerating returns, or AI investment questioned by a second consecutive example of cash flow strain alongside massive capex commitments.

What the Results Could Mean

Bullish scenario: Microsoft reports Azure at the high end of 39-40% guidance with improving operating leverage signals; Copilot seat growth reaccelerates or ARPU metrics show deepening use. Meta reports revenue beating $61 billion with Zuckerberg citing specific agent monetization timelines and early enterprise traction. Both stocks rally. The narrative: AI spending is working, the infrastructure buildout is generating returns faster than feared.

Bearish scenario: Azure guides down toward 37% growth for the next quarter as Alphabet, AWS, and Chinese cloud providers absorb demand; Copilot adoption stalls at 20 million seats. Meta beats on revenue but raises capex again or pushes agent timelines further into 2027. Both stocks fall. The narrative: the capex supercycle is a self-reinforcing arms race with returns perpetually just over the horizon.

Base case: Results near consensus with mixed signals—strong revenue, elevated capex commitments, uncertain forward indicators on monetization. Neither clearly validates nor refutes the thesis. Markets drift, and the real test gets pushed to Q3 earnings in October.

The Bigger Picture

What makes July 29 significant is not just two earnings reports—it is the capstone of a week in which every major technology benchmark gets a real data point. After Alphabet’s negative cash flow, after Intel’s demand validation, after Tesla’s hardware-plus-software progress, the market will use Microsoft and Meta’s results to make a summary judgment: is the AI investment cycle on track to deliver returns, or is it a bet still waiting for its payoff?

The answer matters beyond stock prices. It will influence whether hyperscalers double down on 2027 capex commitments or exercise caution. It will shape how enterprise buyers approach AI procurement budgets in H2. And it will signal to the startup ecosystem whether the capital flows that have been funding AI infrastructure companies are likely to continue.

July 29 is, in the most literal sense, a day of reckoning for the AI spending supercycle. The results are due after market close.