Mag4 week is Cloud and CapEx. Microsoft and Meta print Wednesday; Apple and Amazon Thursday.

Microsoft and Meta report on Wednesday, July 29. Apple and Amazon follow on Thursday, July 30. Alphabet already opened the mega-cap AI window on July 22 — this is the rest of the Mag4 block in one pass.

The frame for the week is not four EPS headlines. It is cloud demand, CapEx, and data-center capacity: who is still accelerating, who already guided a wall of spend, and whether the filed receipts (RPO, AWS growth, CapEx dollars, third-party cloud commitments) still support the build.

Research, not investment advice. Figures below are from SEC-filed Exhibit 99.1 releases and Form 10-Qs unless labeled otherwise (retrieved via Aether / EDGAR).

Calendar

The CapEx and cloud scoreboard (already filed)

That table is the week. Three hyperscaler-style builders and one device/services cash engine that leases and buys compute rather than posting AWS-scale CapEx.

Microsoft — Cloud and RPO first; CapEx is the silent line

Microsoft’s fiscal year ends June 30, so Wednesday is a full-year print, not a mid-year check-in.

Last quarter (Q3 FY2026, ended Mar 31)

From the April 29 Exhibit 99.1:

The demand side is already extreme: Cloud at $54.5B with RPO nearly doubled year over year. The CapEx side shows up cleanly in cash flow — PPE additions roughly doubled versus the year-ago quarter ($16.7B). Guidance lives on the call, not in the press release (“Business Outlook”).

Past year (FY2025, closed Jul 30, 2025)

Trajectory into Wednesday: Cloud grew from a $46.7B Q4 FY2025 run-rate into a $54.5B Q3 FY2026 print; Azure stayed in the high-30s / 40% growth band; PPE spend stepped from ~$65B for FY2025 into an $80B nine-month pace. Fiscal Q4 either extends that or shows the first hard deceleration while CapEx stays hot.

What grades Microsoft on Jul 29

1. Azure growth — hold near Q3’s ~40%, or step down?
2. Microsoft Cloud — still high-20s growth on a larger base?
3. Commercial RPO — still compounding after $627B / +99%?
4. PPE / CapEx commentary — does the FY CapEx path stay consistent with the $30B+/quarter build?
5. AI run-rate — second print after the >$37B claim.

Live multiples: MSFT valuation.

Meta — Ads fund it; CapEx and third-party cloud are the debate

Meta is not a public-cloud landlord. It is an ads machine buying and leasing an enormous compute fleet. The Q1 print already raised the CapEx ceiling mid-year.

Last quarter (Q1 2026)

CFO commentary tied the CapEx raise to higher component pricing and, to a lesser extent, additional data center costs for future-year capacity. The 10-Q adds the contractual footprint: $237.67B of non-cancelable commitments as of March 31, 2026 — “mostly related to third-party cloud capacity arrangements” plus servers, network, data centers, and Reality Labs hardware, with about $42.25B due in 2026 and $47.65B in 2027.

Past year (FY2025)

In one winter Meta went from ~$72B of CapEx to a mid-year guide that can clear $145B. That is the entire Meta AI-infra argument in two filings.

What grades Meta on Jul 29

1. Revenue vs $58–61B guide — and whether ad price × impressions still look like Q1’s +12% / +19% mix.
2. CapEx — still on a $125–145B year, or another raise / mix shift into cloud leases?
3. Operating margin — can ~41% survive infrastructure and third-party cloud costs?
4. Expense path — still inside $162–169B while CapEx runs hot?
5. OI vs 2025 — management still expects 2026 operating income above 2025.

Live multiples: META valuation.

Amazon — AWS re-accelerated; the $200B CapEx year is the backdrop

Amazon is the cleanest “cloud growth vs CapEx burn” twin to Microsoft inside this block.

Last quarter (Q1 2026)

Jassy’s Q1 quote: AWS at 28% was the fastest growth in 15 quarters, on a large base; the chips business topped a $20B revenue run-rate. Capacity color in the same release: OpenAI commitment for ~2 GW of Trainium (ramping 2027); Anthropic up to 5 GW of Trainium generations; 2.1M+ AI chips landed in 12 months. Net income was flattered by $16.8B of pre-tax gains on the Anthropic investment — grade the quarter on AWS and operating income, not EPS.

Past year (FY2025 / Q4 print)

Q1 already beat the company’s own Q1 sales guide range ($173.5–178.5B) and showed AWS accelerating from +20% for FY2025 → +24% in Q4 → +28% in Q1. Thursday asks whether that re-acceleration holds while CapEx stays on a ~$200B annual path.

What grades Amazon on Jul 30

1. AWS growth — still mid-to-high 20s after +28%?
2. AWS operating income — still expanding in dollars while infrastructure spend rises?
3. Sales vs $194–199B and OI vs $20–24B (Prime Day assumed in Q2).
4. CapEx / FCF — is the $200B year still the base case on the call?
5. Ignore Anthropic mark-to-market noise in net income; use segment OI.

Live multiples: AMZN valuation.

Apple — not a hyperscaler CapEx print; Services and the device base are the infra angle

Apple does not report an AWS line. For a Cloud / CapEx / data-center week it is the contrast case: on-device AI, Services (including cloud services inside Services), and a CapEx line that is an order of magnitude smaller than Meta or Amazon.

Last two quarters

Q2 10-Q says Services growth was driven by advertising, the App Store, and cloud services. R&D for the quarter was $11.4B (vs $8.6B) — that is where Apple’s AI spend shows up more than PPE. Apple typically guides on the call, not in the Exhibit 99.1 body.

What grades Apple on Jul 30

1. Services — third straight record, or a pause?
2. iPhone / Greater China — Q2 already showed Greater China at $20.5B; does the install base story hold into the summer quarter?
3. Gross margin — Q2 total GM was 49.3%; Services GM 76.7%.
4. Any AI / Private Cloud Compute capacity color on the call — qualitative, but it is the bridge to the Mag4 CapEx week.
5. Do not grade Apple on hyperscaler CapEx dollars; the comparable is cash generation and Services mix.

Live multiples: AAPL valuation.

One scorecard for Jul 29–30

Wednesday (MSFT + META)

1. Azure / Microsoft Cloud growth and RPO vs the Q3 bar (+40% / +29% / $627B).
2. Meta revenue vs $58–61B and CapEx still inside $125–145B.
3. Whether Meta’s margin can stay near 41% while third-party cloud commitments stay enormous.

Thursday (AMZN + AAPL)

1. AWS growth after +28% — hold or fade — against the ~$200B CapEx year.
2. Amazon OI vs $20–24B guide (clean of Anthropic gains).
3. Apple Services trajectory and any on-device / cloud-services commentary — the non-hyperscaler control sample.

If Microsoft and Amazon both show cloud still accelerating while CapEx guides hold, the AI-infra complex keeps its demand receipt. If cloud decelerates while Meta/Amazon CapEx paths stay maxed, the week becomes a duration-of-spend debate — the same tension Alphabet already put on the table.

Stress-test growth and margin assumptions in the DCF calculator.

Primary sources

Microsoft — Q3 FY2026 Exhibit 99.1 (accession 0001193125-26-191457, Apr 29, 2026); FY2025 / Q4 Exhibit 99.1 (accession 0000950170-25-100226, Jul 30, 2025); Form 10-Q period ended Mar 31, 2026 (accession 0001193125-26-191507).

Meta — Q1 2026 Exhibit 99.1 (accession 0001628280-26-028364, Apr 29, 2026); Q4 & FY2025 Exhibit 99.1 (accession 0001628280-26-003832, Jan 28, 2026); Form 10-Q period ended Mar 31, 2026 (accession 0001628280-26-028526) — CapEx outlook and $237.67B commitments.

Amazon — Q1 2026 Exhibit 99.1 (accession 0001018724-26-000012, Apr 29, 2026); Q4 2025 / FY Exhibit 99.1 (accession 0001018724-26-000002, Feb 5, 2026) — includes ~$200B 2026 CapEx quote; Form 10-Q period ended Mar 31, 2026 (accession 0001018724-26-000014).

Apple — Fiscal Q2 2026 Exhibit 99.1 (accession 0000320193-26-000011, Apr 30, 2026); Fiscal Q1 2026 Exhibit 99.1 (accession 0000320193-26-000005, Jan 29, 2026); Form 10-Q period ended Mar 28, 2026 (accession 0000320193-26-000013).

Filing excerpts retrieved and cross-checked via Aether’s SEC index and EDGAR.

Research, not investment advice.