Wednesday brings Meta and Microsoft, both reporting after the close. Thursday brings Apple, also after the close. Alphabet already set the tone on July 22, raising its 2026 capex guide by $15 billion to $205 billion and watching its stock drop 6% the next day. Between Amazon, Google, Meta, and Microsoft, combined 2026 AI infrastructure spending is tracking toward roughly $725 billion, up about 77% from near $410 billion in 2025. That is the number the week has to answer for. Everything else — ad growth, Azure growth, iPhone units — is downstream of whether the market believes that spend is buying something.
Meta guided second-quarter revenue to $58–61 billion; consensus sits near $60.2 billion with EPS around $7.18–7.23. Last quarter it delivered $56.31 billion, up 33% year over year, the fastest growth since 2021, with adjusted EPS of $7.31 against a $6.79 estimate. The advertising engine is genuinely running at its best pace in years. The problem sitting on top of it is capex: full-year 2026 guidance was raised to $125–145 billion from $115–135 billion, and Meta is the one hyperscaler in the group with no public cloud business to rent that capacity back out through. Amazon, Microsoft, and Google can point to AWS, Azure, and Google Cloud as revenue lines that partially offset the spend. Meta’s buildout serves Meta, full stop, which is why BMO has called it the least visible ROI story among its peers. The stock has already priced in real skepticism — down roughly 20% from its August 2025 high near $790, trading in the $600–640 range into the print, RSI a neutral-ish 46. The 28-analyst six-month median target of $815 implies the sell side hasn’t caught up to that repricing, or hasn’t been forced to yet. Zuckerberg telling employees internally that AI agent progress had lagged expectations didn’t help the case in the weeks before the print.
Microsoft’s setup is a variation on the same theme with a different moat underneath it. Consensus for fiscal Q4 is revenue near $87.7 billion, up about 15%, with EPS around $4.21–4.24. Azure grew roughly 40% last quarter, and the guided range for this print is 39–40% — the single Microsoft number that matters most, because it is the direct offset to the capex argument. AI revenue is running near a $37 billion annualized clip, up 123% year over year. Against that, fiscal 2026 capex guidance stands at $190 billion, and BNP Paribas has floated a fiscal 2027 figure as high as $262 billion. The stock has taken the brunt of it: down about 18% year to date and roughly 25% over twelve months, the weakest of the Magnificent Seven, trading below its 20-, 50-, and 200-day averages heading into Wednesday, at its cheapest forward multiple since 2023. A mean analyst target near $545 sits well above the $381 close, the same gap-to-sentiment pattern showing up in Meta. Layered on top is a securities class action — Levi & Korsinsky and others — alleging misleading statements on Copilot adoption and Azure AI performance between May 2025 and January 2026, with a lead-plaintiff deadline of August 11. Azure’s enterprise lock-in and the OpenAI relationship are real moats; the legal overhang is noise Meta doesn’t carry into its print.
Apple’s Thursday report inverts the pattern entirely, and that is the more interesting story of the week. It is also Tim Cook’s last earnings call as CEO before John Ternus takes over in September and Cook moves to executive chairman. Consensus calls for fiscal Q3 revenue near $108.8–109 billion, up 14–17% off last year’s $94.04 billion, with EPS around $1.88–1.89, up roughly 20% from $1.57. Services is expected near $31.4 billion at better than 70% gross margin, and blended gross margin is guided into the 47.5–48.5% range. Apple’s AI approach has stayed capital-light — on-device Apple Intelligence and a private cloud footprint, not hyperscaler-scale data centers — backed by a multiyear Broadcom custom-silicon commitment worth more than $30 billion aimed at producing upwards of 15 billion U.S.-made chips. The market has rewarded that positioning: the stock closed at $333.02 on July 24, up 22.5% year to date and sitting just below its 52-week high of $334.99 — essentially the mirror image of Meta’s and Microsoft’s charts heading into the same week. What stands out is that the 12-month consensus price target, $318.81, sits below the current price. Analyst models haven’t caught up to the rally, which either means the stock has run ahead of itself or that the market is underwriting a cleaner China and iPhone 17 read than the sell side is willing to put in a spreadsheet yet.
The three moats are different in kind. Meta’s is an attention and ad-monetization engine with no external revenue path for the capital it’s burning. Microsoft’s is Azure and enterprise software stickiness with a legal cloud sitting over the AI-adoption narrative. Apple’s is ecosystem lock-in plus a services annuity, now paired with a capital structure the market is actively rewarding relative to peers spending several multiples more on data centers. The single thing worth watching across Wednesday’s two reports is whether either company shows operating leverage improving alongside the spend, or whether margins keep compressing the way Alphabet’s capex raise implied on July 22. That answer decides whether the gap already priced between the capex-heavy names and Apple’s re-rating widens or closes.