The Magnificent 7: Anatomy of Market Dominance

• Revenue. $96.2B for the quarter ended July 26, 2026, up 106% year on year and 18% on the prior quarter (Nvidia results, 26 Aug 2026, as reported by iPhone in Canada and Plus500).
• Data Center. $89.0B, up 117% year on year.
• Margin. Gross margin 75.0%, GAAP and non-GAAP; operating income $63.7B.
• Outlook. Q3 guidance of about $108B, plus or minus 2%, with no Data Center compute revenue from China assumed.

• Total. Q3 2026 capex across five spenders was about $227B, up about 20% from $188.5B in Q2; the 2026 forecast is $803B to $828B (Business Insider via Hargreaves Lansdown, 30 Oct 2026).
• Amazon. Q3 cash capex $75.9B; about $220B guided for 2026, maintained, with meaningfully more expected in 2027.
• Microsoft. About $170B of calendar-year capex, below its prior outlook.
• Meta. 2026 range reported as $135-145B (press reports of the Q3 call; not independently confirmed here).

Schwab, as summarised in a search result of Aug 2026, put the combined market value of the seven at nearly $24 trillion, about 34% of a roughly $70 trillion S&P 500, with Nvidia near $5.4 trillion. Other 2026 readings ranged from about 33% (Motley Fool, July 2026) to 35.7% (Rallies.ai, after earnings); one blog claim of about 45% conflicts with all of these and is not used.

Amazon's AWS reported about $46B of revenue for the September 2026 quarter, up about 39% year on year, its fastest growth in nearly five years (Techzine, Oct 2026). Press reports also cite Azure growth of 48% for the same quarter; this was not confirmed from a primary source here.

Not refreshed: national equity market caps (June 2024), 2024 return comparison, 5-year betas, FY2024 revenue and gross margin charts for the other six companies, GPU ownership (EpochAI, June 2024), cloud market share (SRG, 2025), Meta, Apple and Tesla current-quarter figures, antitrust case status.

1. 1976-2004Founded: the seven span three tech eras, Apple through Tesla and Meta.
2. 2007Mobile: the iPhone mints Apple's global scale and cash machine.
3. 2012Deep learning: AlexNet on Nvidia GPUs lights the AI fuse.
4. ~2015~12% of the index: already big, not yet dominant.
5. 2020Digital surge: COVID and passive flows accelerate concentration.
6. 2023-24The AI boom: the seven reach ~35% of the S&P 500.

Seven companies, seven completely different machines: an ecosystem, a cloud, an ad auction, a chip monopoly, a search box, a store, and a car maker. What unites them is scale, cash generation, and now a shared dependence on the AI trade.

• Azure. ~20% cloud share, ~$75B run-rate, still growing ~39% a year with AI.
• Microsoft 365 + Office. the productivity default across the corporate world.
• Entra ID. ~600M identities: switching means ripping out your log-in backbone.
• Backlog. ~$368B of committed future revenue underwrites the capex.

• Net sales. $638B in FY2024, the largest revenue base in the M7.
• AWS. ~$108B of that revenue, but historically ~74% of operating profit.
• Flywheel. third-party sellers add selection without adding inventory.
• Custom silicon. Trainium and Inferentia chips cut its Nvidia bill.

• Reach. 3.4 billion daily users across Facebook, Instagram, WhatsApp and Threads.
• Ads. 98% of $165B revenue; ~23% of global digital ad spend.
• Margins. 82% gross, 42% operating: not competed away in 14 years.
• The bet. $69B of cumulative Reality Labs losses and a $65B+ AI capex plan.

1. Hardware & commerce: iPhones, EVs and retail, huge scale, thinner margins.
2. Software & services: Office, iOS services and subscriptions, sticky and high-margin.
3. Cloud: AWS, Azure and Google Cloud rent compute as a metered, recurring utility.
4. Advertising: Google and Meta together capture roughly half of all global digital-ad dollars.

The Magnificent 7 and their peers are pouring a very large and rising sum into AI infrastructure (a combined forecast of $803B to $828B for 2026 across Amazon, Microsoft, Alphabet, Meta and SpaceX, per Business Insider via Hargreaves Lansdown, Oct 2026; more than $300 billion a year when this piece was first written): the largest, fastest capital build-out in corporate history. It is also their single biggest shared bet: if AI demand disappoints, they disappoint together.

1. Apps: Copilot, Meta AI, Gemini, Alexa+, where users actually pay.
2. Models: OpenAI (Microsoft), Gemini (Google), Llama (Meta), Claude via AWS/Azure.
3. Cloud: AWS, Azure and Google Cloud rent the compute as a utility.
4. Chips: Nvidia GPUs, plus in-house Trainium, TPU and Maia silicon.

1. AI optimism lifts equities.
2. Cash pours into index ETFs.
3. M7 are 30-40% by weight.
4. Buying lifts M7 prices.
5. Their index weight rises further, and the reflexive loop repeats.

1. Macro shock: a rate or growth surprise triggers a risk-off move.
2. Redemptions: investors sell index funds, actively or passively.
3. Forced selling: the M7's heavy weight means they are sold hardest.
4. Correlations spike: diversification fails exactly when you need it.

A two-axis map plots each company by profit-engine durability (vertical) against dependence on the AI-capex cycle (horizontal). Apple sits in the cash-rich, AI-light corner. Meta, Microsoft, Alphabet and Amazon sit in the cash-rich, AI-levered quadrant, with Nvidia furthest along the AI-dependence axis. Tesla sits lower, toward the fragile, AI-exposed side.

• Concentration. ~35% of the index in seven correlated names.
• AI capex. $300B+ a year in 2025, with 2026 forecasts above $800B across the big spenders, that must eventually earn its return.
• Valuation. priced for durable dominance and pricing power.
• Regulation. antitrust in the US and EU targets several at once.
• Geopolitics. Nvidia's China sales curbed by export controls.
• Reflexivity. passive flows amplify both the rise and the fall.

Culper Research, a short seller, posted on 13 May 2026 that it is short Nvidia and gains if the price falls. Its report, 'NVIDIA (NVDA): The China Problem', says over 20% of FY2026 compute revenue stayed China-driven through illegal diversion and Southeast Asian intermediaries, and that documents suggest Alibaba financed Megaspeed, a Singapore buyer, through a shell company (Culper Research, 13 May 2026; Benzinga, May 2026). I could not open it (HTTP 403), so this is secondhand.

I found no reply to the report by name. Nvidia's 10-Q says it was 'effectively foreclosed' from China's data center compute market and that it relies on customers' compliance programs after sale (Nvidia, 10-Q, 26 Aug 2026). Nvidia has said its Megaspeed site visits gave no reason to believe products were diverted (CNBC, 10 Oct 2025) and that diversion claims about Megaspeed were 'thoroughly investigated and debunked last year' (The Hill, 29 Oct 2026). The March 2026 indictment Culper cites names neither Nvidia nor Megaspeed (US Department of Justice, 19 Mar 2026). I found no court ruling or Federal Register notice naming Megaspeed (US Federal Register, 11 Nov 2026).

Separately, Nvidia's board announced an independent investigation on 8 November 2026 after reports that its CEO was the target of a foreign intelligence effort. The release names no other company and EDGAR shows no Form 8-K for it; I found no source tying it to Culper's report, and press reports, not the release, link that reporting to Megaspeed (Nvidia, 8 Nov 2026; SEC EDGAR, 11 Nov 2026; The Hill, 29 Oct 2026). Three House committee chairs requested a closed-door briefing for 18 November (House Select Committee on China, 5 Nov 2026).

Owning the market used to mean owning America's breadth. Today it increasingly means a concentrated bet on seven companies, and on the AI trade that lifts them together. That may prove a very good bet. It is simply no longer a diversified one.

1. Does AI capex earn its return? ROIC versus WACC across the hyperscalers.
2. Does concentration mean-revert? Or push past 35% of the index.
3. Do the antitrust cases bite? Google, Amazon, Apple and Meta all in the dock.
4. Can Nvidia hold 75% margins? The clearest read on AI pricing power.
5. Do passive flows reverse? The flywheel spins both ways.

The S&P 500 still lists five hundred companies. But a third of it now beats with seven hearts, and when the AI trade breathes in, they all inhale together.

1. Seven stocks are the market: about a third of the S&P 500 (~34% in Aug 2026, ~35% at the 2024 peak).
2. The moats are real but different: and each has a distinct soft spot.
3. AI capex is the shared bet: hundreds of billions a year (2026 forecasts above $800B) bind their fates together.
4. Diversification has quietly eroded: the index moves with the seven.
5. The question is durability vs price: extraordinary businesses, demanding valuations.