Meta in the AI stack

The business and what the filings show

Meta is the large buyer that doesn’t rent out what it builds. It spent $104B over its latest four quarters, 78% of its operating cash flow, to run AI across its own apps: ranking, recommendations, advertising tools and its own models. Its return has to show up in advertising revenue and engagement rather than in a cloud segment, which makes it harder to see from outside.

Revenue over the latest four quarters was $242B, +28%, at an 81% gross margin. Meta’s commitments go well beyond capex: it reports hundreds of billions of dollars of non-cancelable contractual commitments, most for third-party cloud capacity, leases signed but not yet started, and two large data-center campuses held in joint ventures it doesn’t consolidate, backed by residual value guarantees.

That structure keeps much of the build-out off Meta’s balance sheet, which is a financing choice rather than a problem in itself. The question is whether advertising growth keeps pace with the fixed commitments. Meta has also bought capacity from CoreWeave and agreed nuclear power contracts with Vistra, and AMD has issued it a warrant tied to accelerator purchases.

How AI shows up in the numbers

AI is embedded across Meta’s products rather than sold as a separate line: it decides which posts and ads people see, generates ad creative, and powers Meta’s own assistant and models. The filings show it as capex, as third-party cloud commitments and in advertising growth, which is the only revenue line that has to repay the spending.

Meta’s $104B of capital spending over the latest four quarters is 18% of the $566B the five largest buyers spent over their latest four quarters. The five together raised capex +81% on the four quarters before; Meta’s own revenue grew +28%.

How it talks about AI

In its latest earnings release, furnished to the SEC on October 21, 2026, Meta mentioned AI 5 times in 1,469 words of narrative text, about 3.4 mentions per 1,000 words. Of its 5 sentences about AI, 1 included a number: a revenue figure, a growth rate, a customer count or another measurable result. Across the 8 releases we read, back to January 29, 2025, AI mentions per 1,000 words went from 5.1 to 3.4.

Specific claims are easier to check against the numbers than general enthusiasm, which is why the share of AI sentences that carry a figure is worth watching alongside how often AI is mentioned. The sentences below are quoted from the release as published, after removing safe-harbor text and company boilerplate.

• “Separately, we remain focused on improving Reality Labs profitability and expect losses to narrow next year to around $15 billion as we reallocate investment toward other priority areas, primarily AI. 2 Establishment of New Class C Common Stock The Board of Directors has approved the establishment of a new class of…”
• “Looking beyond 2026, we are still developing our budget and capacity plans for next year, but we expect to continue investing aggressively to support both our core business and emerging AI initiatives.”
• “Our AI investments are already delivering substantial incremental advertising revenue through stronger engagement, better recommendations, and more efficient monetization.”

Source: earnings release (Form 8-K, Exhibit 99.1), October 21, 2026.

The trend in the numbers

Since the start of 2023, Meta’s largest quarter by revenue was Q3 2026 at $64.6B, and its highest gross margin was 82% in Q2 2025. The latest quarter set the revenue record for the period.

Quarter by quarter, Meta’s revenue grew +24% in Q4 2025, +33% in Q1 2026, +28% in Q2 2026, +26% in Q3 2026 on the same quarter a year earlier. The pace has been steady.

Gross margin in the latest quarter was 79%, against 82% a year earlier, so more of each new dollar is passing through to costs. Operating margin moved from 40% to 20% over the same span. Over the latest four quarters it turned 55% of its revenue into operating cash flow and spent 43% of revenue on capital projects, against 33% a year earlier.

Where it sits in the stack

Meta sits in the AI clouds layer of the AI stack. It is paid by the software layer; its own spending flows to servers, networking, power & DCs, chips.

Among the 8 companies in this layer on our map, it ranks number 4 by revenue and number 3 of 7 by revenue growth. Its gross margin of 81% compares with 61% for the layer as a whole, a sign that it keeps more of each dollar than its neighbors. The full layer analysis is on the AI clouds page.

Latest four quarters for 10-Q filers; latest fiscal year for 20-F filers. Companies under $100 million of revenue are left out.

Against the 7 US-listed companies in its layer, Meta grew +28% over the latest year versus +20% for the layer combined, and converted 12% of revenue into free cash flow versus 3% for the layer. It is taking share of the layer’s revenue.

What has to go right, and what could go wrong

What has to go right. AI has to keep improving ad targeting and engagement enough for advertising revenue to grow into the commitments, and Meta’s own models have to reduce what it pays third-party clouds over time.

What could go wrong. Meta has no cloud business to sell spare capacity through, so if the return through advertising disappoints, the leases, cloud contracts and venture guarantees remain fixed costs.

These describe the conditions the business depends on, as the filings show them. They are not a view on the stock.

Through the lens of Investing in AI

Rob May’s book offers six mental models for telling AI moats from traps. The one most useful for Meta’s layer is “Expect major market dislocations”: Most markets move slowly. AI does not. A single technical result can force the market to reprice an entire layer overnight. Assume fast repricings and sudden business-model breaks are normal, and position to benefit from shocks.

The question it asks of any company is: Which three to five assumptions have to hold for this investment to work? Applied to Meta, the filings on this page are the place to start: the margin trend shows whether it holds scarce value or passes it through, the customer and financing disclosures show who it depends on, and the commitments show what it has promised before the revenue arrives.

Customers and concentration

US companies must disclose any customer that accounts for 10% or more of revenue, and foreign filers usually describe their largest customers. Concentration matters because a company’s growth is only as durable as the handful of buyers behind it, and in the AI stack those buyers are often the same few companies.

• “No customer represented 10% or more of total revenue or accounts receivable for the years ended December 31” 10-K, January 29, 2026

Financing links

Investments, guarantees, purchase commitments, warrants and power contracts between Meta and other companies, each quoted from the filing that discloses it. Numbers match the money-loop map.

• 17Meta → CoreWeavePurchase commitment

  Up to $21.0B through 2032

  “Similarly, in March 2026, we entered into an order form under an existing master services agreement pursuant to which Meta Platforms, Inc. ("Meta") initially committed to pay us up to approximately $21.0 billion (inclusive of access to new computing capacity through December 20, 2032 and the exercise of an existing option to access additional computing capacity through April 10, 2032) and, as a result, we also expect Meta to be a significant customer in future periods.”

  CoreWeave 10-Q for the period ended June 30, 2026, filed August 12, 2026

• 20AMD → MetaCustomer warrant

  Warrant for up to 160M AMD shares at $0.01

  “In October 2025 and February 2026, the Company issued warrants to OpenAI OpCo, LLC (OpenAI) and Meta Platforms, Inc. (Meta), each entitling the holder the right to purchase up to 160 million shares of the Company’s common stock at an exercise price of $0.01 per share (the OpenAI Warrant and the Meta Warrant, respectively).”

  AMD 10-Q for the period ended June 27, 2026, filed August 5, 2026

• 22Meta → Meta’s data-center venturesCredit or guarantee

  $41.5B of residual value guarantees, $68.4B maximum exposure

  “The following table summarizes our maximum exposure to loss in the unconsolidated data center ventures (in millions): September 30, 2026 December 31, 2025 Carrying value of equity investments $4,732 $1,826 Lease commitments 18,476 12,311 Future funding commitments 3,703 3,724 Maximum residual value guarantees 41,493 28,091 Total $68,404 $45,952 Other In addition, we have other types of unconsolidated VIEs of which we are not the primary beneficiary.”

  Meta 10-Q for the period ended September 30, 2026, filed October 22, 2026

• 23Meta → Third-party cloudsPurchase commitment

  $454.5B of non-cancelable commitments, mostly cloud capacity

  “As of September 30, 2026, we had $454.52 billion of non-cancelable contractual commitments, comprising both short-term and long-term arrangements, most of which are related to third-party cloud capacity arrangements and other investments in technical infrastructure, and we continue to enter into additional significant contractual arrangements.”

  Meta 10-Q for the period ended September 30, 2026, filed October 22, 2026

• 25Meta → VistraPower contract

  2,609 MW of nuclear power under 20-year PPAs

  “In January 2026, Vistra announced it had entered into 20-year PPAs with Meta, pursuant to which the Company has agreed to supply Meta with a total of 2,609 MW of carbon-free power and capacity from the Company's PJM nuclear power plants as follows: 1,268 MW of energy and capacity from Perry and 908 MW of energy and capacity from Davis-Besse; and 213 MW of uprate energy and capacity from Perry, 80 MW of uprate energy and capacity from Davis-Besse, and 140 MW of uprate energy and capacity from Beaver Valley.”

  Vistra 10-Q for the period ended June 30, 2026, filed August 10, 2026

Commitments and backlog

• Backlog. Meta doesn’t report a comparable total of remaining performance obligations in its XBRL data.
• Purchase commitments. $356B of unconditional purchase obligations as of September 30, 2026 (10-Q): what it has contracted to buy from its own suppliers.
• Server life. Meta depreciates its servers and network equipment over 5 to 5.5 years, after changes it disclosed: 5→5.5 years (fiscal year 2025). See the useful-life calculator.
• Leases signed, not yet started. “These lease obligations were approximately $355.73 billion, consisting of data centers, energy-related infrastructure, colocations, and certain network infrastructure, which will commence during the remainder of 2026 through 2036 with lease terms ranging from greater than one year to 30 years.” 10-Q

What to watch

Meta’s next quarterly report will cover the quarter ending around December 30, 2026. Large companies must file a 10-Q within 40 days of a quarter’s end and a 10-K within 60 days of the fiscal year’s end, so the new figures should arrive by about February 28, 2027 at the latest. This page, the layer pages and the next quarterly report all update from that filing.

• Advertising revenue growth against capex growth: advertising is the only revenue line that has to repay Meta’s spending.
• Non-cancelable commitments and the joint-venture guarantees, the fixed costs that sit outside the balance sheet.
• Purchases from third-party clouds such as CoreWeave, which show how much capacity Meta rents rather than builds.
• Its place in the next quarterly AI Stack Report, which updates every company on the map from the new filings.

Figures from Meta’s SEC filings, latest period ended September 30, 2026. Other companies: all company pages.