AI REVENUE WARS ● 2026-10-30

Explorer

Findings

Generated from the dataset at build time (comparison quarter 2026Q2). The same text is written to REPORT.md. Numbers are filed revenue unless labelled; nothing here claims causation.

Strongest revenue growth

Fastest trailing-twelve-month (TTM) revenue growth at 2026Q2 among companies and segments with at least $1B TTM revenue, and the largest dollar gains.

  • CoreWeave: TTM $7.6B, +115.3% vs a year earlier.
  • NVIDIA Data Center: TTM $277.8B, +89.5% vs a year earlier.
  • Reddit: TTM $2.8B, +66.6% vs a year earlier.
  • Google Cloud: TTM $77.6B, +57.8% vs a year earlier.
  • Figma: TTM $1.3B, +43.4% vs a year earlier.
  • Cloudflare: TTM $2.5B, +33.5% vs a year earlier.
  • Snowflake: TTM $5.4B, +32.0% vs a year earlier.
  • Datadog: TTM $4.0B, +31.5% vs a year earlier.
  • Largest incremental TTM revenue: NVIDIA Data Center $131.2B; Meta advertising $48.0B; Google Search & other $35.0B; AWS $32.0B; Google Cloud $28.4B; Amazon Ads $14.8B; Oracle $10.0B; YouTube ads $4.5B.
What the data shows
4 of the 8 fastest growers and 3 of the 8 largest dollar gains are chip, cloud or GPU-cloud businesses; the rest are advertising and software companies.
What it suggests
Through 2026, AI-related spending is concentrated in infrastructure. Revenue at the application layer grows at rates close to its pre-2023 trend.
What remains unknown
How much of the infrastructure revenue rests on end-customer demand vs capacity built ahead of demand with venture or debt funding; one company's revenue here is often another's capital spending.

Acceleration and deceleration

Acceleration = YoY growth in 2026Q2 minus YoY growth in the same quarter a year earlier, in percentage points (companies with at least $500M TTM revenue).

  • Accelerating most: NVIDIA Data Center +56.4% → +116.6% (+60.2 pp); Google Cloud +31.7% → +81.8% (+50.1 pp); AWS +17.5% → +36.8% (+19.3 pp); DigitalOcean +13.6% → +28.6% (+14.9 pp); Fastly +12.3% → +23.3% (+10.9 pp); Snap +8.8% → +18.9% (+10.1 pp).
  • Decelerating most: CoreWeave +206.8% → +112.5% (-94.3 pp); Reddit +77.7% → +61.1% (-16.6 pp); The Trade Desk +18.7% → +3.0% (-15.7 pp); GitLab +29.2% → +21.3% (-7.9 pp); Klaviyo +31.9% → +26.4% (-5.5 pp); Elastic +19.5% → +15.1% (-4.4 pp).
What the data shows
Infrastructure businesses among the six strongest accelerations: NVIDIA Data Center, Google Cloud, AWS, DigitalOcean, Fastly. The six largest decelerations started from YoY growth of 19–207%.
What it suggests
The re-acceleration in cloud is consistent with AI workloads. Deceleration at very high growth rates (CoreWeave, AppLovin, Reddit) is partly mechanical, because the base keeps growing.
What remains unknown
Company-specific drivers (pricing, contract timing, divestitures) are not separated here; the AI-attributed share of each company's growth is not disclosed consistently.

Slowdowns and low growth

Lowest three-year compound growth in TTM revenue (2023Q2 to 2026Q2, companies with at least $1B TTM three years ago).

  • Teradata: $1.8B → $1.7B, -2.2% a year.
  • ZoomInfo: $1.2B → $1.3B, 1.6% a year.
  • Akamai: $3.7B → $4.3B, 5.7% a year.
  • GoDaddy: $4.2B → $5.1B, 7.1% a year.
  • Salesforce: $33.1B → $43.9B, 9.9% a year.
  • Oracle: $50.0B → $67.4B, 10.5% a year.
  • Adobe: $18.4B → $25.2B, 11.0% a year.
  • Twilio: $4.1B → $5.6B, 11.2% a year.
What the data shows
5 established vendors with over $1B of revenue grew less than 10% a year over three years; Teradata shrank.
What it suggests
Some of these (sales data, web presence, data warehousing) are in categories where AI-native competitors are commonly described as substitutes. That makes them candidates for displacement, nothing more.
What remains unknown
Slow growth has many causes (pricing, macro, seat reductions, company execution). No customer-level evidence in this dataset yet links any of these slowdowns to AI products.

Market expansion

Matched-panel TTM revenue added between 2025Q2 and 2026Q2, by market (public companies with filed revenue in both periods).

  • AI compute & GPU cloud: $135.3B added (+90.2%; 2 companies).
  • Advertising & search: $104.8B added (+21.2%; a year earlier +15.4%; 7 companies).
  • Cloud infrastructure: $61.6B added (+35.6%; a year earlier +21.1%; 6 companies).
  • Databases & data platforms: $12.1B added (+18.1%; a year earlier +9.8%; 5 companies).
  • CRM & go-to-market: $5.6B added (+12.3%; a year earlier +9.3%; 5 companies).
  • Design & creative: $3.0B added (+12.7%; 2 companies).
  • Developer tools & coding: $1.7B added (+25.6%; a year earlier +21.0%; 3 companies).
  • Observability: $1.3B added (+26.5%; a year earlier +23.2%; 2 companies).
  • Customer support & contact center: $1.1B added (+16.1%; a year earlier +10.3%; 3 companies).
What the data shows
AI compute and cloud panels added $196.9B; the six application and data software panels added $24.7B combined; advertising added $104.8B.
What it suggests
Most of the new money measured so far is infrastructure expansion and incremental spending. Budget moving between software vendors would not create net new revenue on this scale.
What remains unknown
Panels exclude private companies, so expansion captured by private AI-native vendors is invisible here and appears only as ARR or run-rate disclosures.

Competitive displacement: what can and cannot be shown

Correlated growth rates are not evidence of displacement. Displacement is recorded only from sourced statements (customer migrations, company disclosures). That evidence is still being collected, so this section lists measurements and testable hypotheses, not conclusions.

  • Share of the measured ad panel, 2023Q4 → 2026Q2: Meta advertising 33.5% → 37.2%; Amazon Ads 11.9% → 12.7%; Reddit 0.2% → 0.5%; Pinterest 0.8% → 0.8%; Snap 1.2% → 1.1%; YouTube ads 8.0% → 7.1%; Google Search & other 44.4% → 40.7%. Every member's dollars grew; the shifts are relative.
  • Hypothesis to test: AI coding tools vs incumbent dev tools. GitLab YoY growth went from +29.2% to +21.3%, while Cognition reports fast ARR growth (Cursor pending). No migration evidence is recorded yet.
  • Hypothesis to test: AI sales-data tools vs ZoomInfo. ZoomInfo grew +1.2% YoY in 2026Q2. Clay, Apollo and Attio disclosures are pending.
  • Hypothesis to test: AI support agents vs contact-center software. Five9 grew +10.3% YoY; Sierra and Decagon report growth from small bases. Zendesk is private.
What the data shows
Within the advertising panel, the largest share gain was Meta advertising (+3.7 pp) and the largest loss Google Search & other (-3.8 pp). No direct displacement evidence is recorded yet.
What it suggests
Displacement is plausible in sales data, support and web building, but the measured dollars are small next to infrastructure expansion.
What remains unknown
Whether any specific budget moved from one vendor to another. That requires the sourced customer and company evidence being collected.

Public versus private

Private companies disclose ARR or annualized run-rate at a point in time; public companies file recognized revenue for a period. The two are listed side by side and never summed.

  • Anthropic: $65.0B run-rate (as of 2026-07-31, reported).
  • OpenAI: $50.0B run-rate (as of 2026-09-30, reported).
  • Databricks: $7.0B run-rate (as of 2026-08-13, reported).
  • Cursor (Anysphere): $4.0B run-rate (as of 2026-06-08, reported).
  • Canva: $4.0B ARR (as of 2025-12-31, reported).
  • Cognition: $1.0B run-rate (as of 2026-09-25, reported).
  • Fireworks AI: $1.0B run-rate (as of 2026-07-15, reported).
  • Vercel: $600M run-rate (as of 2026-10-01, reported).
  • Lovable: $600M run-rate (as of 2026-09-24, reported).
  • Grafana Labs: $400M ARR (as of 2025-09-30, reported).
  • Modal: $300M run-rate (as of 2026-10-14, reported).
  • Glean: $300M ARR (as of 2026-05-28, reported).
  • Fivetran: $300M ARR (as of 2024-10-11, reported).
  • Perplexity: $250M run-rate (as of 2026-08-24, reported).
  • ClickHouse: $250M run-rate (as of 2026-05-27, reported).
  • Zendesk: $250M ARR (as of 2026-05-25, reported).
  • Cohere: $240M ARR (as of 2025-12-31, reported).
  • Sierra: $200M ARR (as of 2026-05-31, reported).
  • Chronosphere: $160M ARR (as of 2025-09-30, reported).
  • Replit: $150M run-rate (as of 2025-09-10, reported).
  • Decagon: $100M run-rate (as of 2026-10-13, reported).
  • Crusoe: $100M ARR (as of 2026-09-17, reported).
  • Intercom: $100M ARR (as of 2026-05-14, reported).
  • Clay: $100M ARR (as of 2025-12-08, reported).
  • Gamma: $100M ARR (as of 2025-11-10, reported).
  • Sentry: $100M ARR (as of 2024-05-15, reported).
  • Webflow: $100M ARR (as of 2022-03-16, reported).
  • Framer: $50M ARR (as of 2025-08-28, reported).
  • Bolt (StackBlitz): $40M ARR (as of 2025-03-03, reported).
  • Supabase: $20M ARR (as of 2024-12-31, reported).
  • Anthropic: $16.8B quarterly revenue for 2026Q3 as reported by Van Data Team (Reuters/FT via attendees) (reported; not a filing).
  • xAI: $2.6B quarterly revenue for 2026Q2 as reported by Safeguard (SpaceX offering document summaries) (reported; not a filing).
  • Anthropic: $16.3B quarterly revenue for 2026Q2 as reported by SEC EDGAR (verified; not a filing).
  • xAI: $818M quarterly revenue for 2026Q1 as reported by Safeguard (SpaceX offering document summaries) (reported; not a filing).
  • xAI: $3.2B quarterly revenue for 2025Q4 as reported by Xix (SpaceX S-1 summaries) (reported; not a filing).
  • Replit: $240M quarterly revenue for 2025Q4 as reported by Entrepreneur Loop (reported; not a filing).
  • Canva: $3.5B quarterly revenue for 2025Q4 as reported by Canva (reported; not a filing).
  • xAI: $107M quarterly revenue for 2025Q3 as reported by Okara (Bloomberg) (reported; not a filing).
  • Lambda: $520M quarterly revenue for 2025Q3 as reported by Forge (reported; not a filing).
  • Lambda: $250M quarterly revenue for 2025Q2 as reported by Forge (reported; not a filing).
  • Anthropic: $1.2B quarterly revenue for 2025Q2 as reported by SEC EDGAR (verified; not a filing).
  • xAI: $52M quarterly revenue for 2025Q1 as reported by Sherwood (Bloomberg) (reported; not a filing).
  • xAI: $2.6B quarterly revenue for 2024Q4 as reported by Xix (SpaceX S-1 summaries) (reported; not a filing).
  • No sourced disclosure yet for 7 of 39 private companies: Mistral AI, Together AI, Baseten, Neon, Pinecone, Attio, Apollo.io.
What the data shows
The largest private disclosure is Anthropic at $65.0B (annualized); for scale, Salesforce's filed TTM revenue is $43.9B and Adobe's $25.2B.
What it suggests
A meaningful share of new AI application spending goes to private companies that public panels cannot see.
What remains unknown
Recognized revenue, gross margins and churn for nearly all private companies; how much of their revenue is pass-through compute cost.

Data limits

What this dataset cannot show yet.

  • Segment definitions change. Microsoft recast its segments twice and only recently began tagging Azure in dollars; Salesforce replaced its per-cloud lines with Agentforce Apps and Data 360. Series stop rather than splice.
  • Fourth quarters are mostly derived as fiscal-year revenue minus nine months, and are marked as derived.
  • Values are as first reported. Later restatements (divestitures, discontinued operations) are flagged, not merged.
  • Foreign filers (Wix, NICE, Nebius) provide annual figures only in SEC XBRL, so they are missing from quarterly panels.
  • AI-specific revenue, customer counts and retention come from earnings releases and are still being added.