Skip to content
Slow Variables
Ask

Model

Lab vertical integration (acquisitions and acquihires, trailing four quarters)

Deals in the trailing four quarters by which an AI lab or a compute company took a company in the layers above the model layer (training input, serving and orchestration, deployment and application), by purchase, acquihire, or a licence or stake that came with its leaders; derived metric lab_vertical_integration_events_4q over the acquisition ledger (acq.<acquirer>.<target>.pt rows, each with a fetched announcement), searched under a written rule and read only for quarters a search has covered.

Reason through this

The reading

A search of every counted buyer's deals since January 2025, under a rule written down first, more than tripled the count in September 2026: the labs, OpenAI above all, and the compute companies are taking coding tools, evaluation and inference software, agents and consumer apps at a steady clip since late 2025, after a slower first half of 2025. The trailing year is read only from the end of 2025, and the direction rule waits for five quarters a search has covered.

Every readingeach dot links to its record
count
28 · 2025-12-31 · reported40 · 2026-03-31 · reported46 · 2026-06-30 · reported
a readingthe rule needs 5 readings; 3 so far
reported

3 readings drawn from the derived metric lab_vertical_integration_events_4q; 75 observations in the evidence.

Derived (lab_vertical_integration_events_4q) from: SEC EDGAR filings (8-K, 6-K, 10-Q exhibits); Anthropic newsroom; CNBC; TechCrunch; OpenAI research posts; CoreWeave newsroom; AMD blog; Calcalist Ctech; Crusoe newsroom; Emmi AI news; Nicolas Bustamante's blog (Fintool founder); Google blog (The Keyword); Gretel blog; Humanloop website; Koyeb blog; Limitless website; Manus blog; Nebius newsroom; Poolside blog; Silicon Republic; Suno blog; Workshop Labs blog; AWS Contact Center Blog; Business Insider; Cirrus Labs website; Jams website; Weights website; BetaKit; Deepgram announcements; Rebellions announcements. SEC EDGAR archives. Anthropic newsroom. CNBC. TechCrunch. OpenAI. CoreWeave (company announcement). AMD blog. Calcalist Ctech. Crusoe newsroom. Emmi AI news. Nicolas Bustamante's blog (Fintool founder). Google blog (The Keyword). Gretel blog. Humanloop website. Koyeb blog. Limitless website. Manus blog. Nebius newsroom. Poolside blog. Silicon Republic. Suno blog. Workshop Labs blog. AWS Contact Center Blog. Business Insider. Cirrus Labs website. Jams website. Weights website. BetaKit. Deepgram (company announcement). Rebellions (company announcement).

The numbers
as ofreadingvaluerecord
2026-06-30—46reported
2026-03-31—40reported
2025-12-31—28reported

Why it matters

The application layer's exit bell. When the labs buy the layer above them, the surplus that layer was capturing moves down the stack; a rising count is the capture thesis's clearest leading signal.

What cuts against it

The ledger holds what parties and filings confirm, so quiet deals are missed until they surface, and acquihires surface late; a rise can reflect how thoroughly a quarter was searched. Many deals are small team hires that say little about margins, and one buyer's spree (OpenAI's) moves the count on its own.

How it is measured

Four quarters. A yearly count of rare events varies between two years with no trend by about the square root of twice its level, which at the December 2025 count after the September 2026 search is about seven, so a smaller move is stable. More deals per year concentrate the application layer into the labs.

Applied to metric:lab_vertical_integration_events_4q.

Derived rows (3)
as ofdimsvalueinputs
2026-06-3046sources123+43 more
2026-03-3140sources123+38 more
2025-12-3128sources123+27 more
Latest readings
  • 2026-06-30 · 46as of 2026-06-30(46 obs)
  • 2026-03-31 · 40as of 2026-03-31(41 obs)
  • 2025-12-31 · 28as of 2025-12-31(30 obs)
Why this status, and its history

It has 3 readings and its direction rule compares 5, so it reads emerging rather than not yet measurable: the site uses not yet measurable only for what cannot be measured at all. The status turns into a direction once enough readings exist.

  1. 2026-09-27not yet measurable to emergingconf 55 → 55 · Claude (model)

    It has 3 readings and its direction rule compares 5, so it reads emerging rather than not yet measurable: the site uses not yet measurable only for what cannot be measured at all. The status turns into a direction once enough readings exist.

    Weighed against it: None needed: no reading has changed; this corrects the word used for too few readings.

    sources123456+37 more
  2. 2026-09-23not yet measurable to not yet measurableconf 55 → 55 · Claude (model)

    The ledger was searched on 23 Sep 2026 under a rule written down first, for all 120 counted buyers from January 2025, and an independent review re-searched the heaviest buyers: it gained 49 deals and lost two that fall outside its scope (TBPN, a media company, and DuckLabs, a database maker). The count now reads 27 at 31 Dec 2025, 38 at 31 Mar 2026 and 43 at 30 Jun 2026. The direction rule compares five quarters a search has covered, so it cannot be read until the search after the end of 2026.

    Weighed against it: The count is still a floor: four deals met the rule without a citable page, and acquihires surface late, so later searches can add deals to quarters already read. The rise across the three readings is mostly the slow first half of 2025 leaving the window; since late 2025 the quarters have held about a dozen deals each. The dead band is now seven deals, about the spread between two years' counts with no trend.

    sources123456+37 more
  3. 2026-09-23concentrating to not yet measurableconf 55 → 55 · Claude (model)

    The count now reads only completed quarters whose trailing year lies wholly after 1 January 2025, when the ledger begins: 8 at 31 Dec 2025, 10 at 31 Mar 2026 and 11 at 30 Jun 2026. The direction rule compares five such quarters, so it cannot be read until the end of 2026. The earlier reading of concentrating compared counts of 3, 5, 6, 9 and 9, the first two from windows that began before the ledger did. The check of 22 Sep 2026 also found two OpenAI purchases missing (Sky in October 2025, Neptune in December 2025) and Weights & Biases dated at its close rather than its announcement; all three are now in the ledger.

    Weighed against it: The ledger is kept by hand and has gaps, so its count is a floor; even five quarters inside it will compare a thinner 2025 with a fuller 2026 unless the ledger is searched again under a written rule.

    sources123456+5 more
  4. 2026-09-22unclear to concentratingconf 55 → 55 · Claude (model)

    The rule's window is now the five complete quarters from the second quarter of 2025 to the second quarter of 2026, after the metrics runner stopped emitting the quarter still under way (it had been dated 30 Sep 2026 and counted while partial). Trailing-four-quarter deal counts over that window: 3, 5, 6, 9 and 9. The count rose by 6, past the dead band of 1, and three of the four steps rose, so the direction rule reads concentrating.

    Weighed against it: The tally is kept by hand from fetched announcements, so it undercounts deals never announced and can rise with disclosure as much as with buying; the latest quarter added no deal; and the reading moved because a partial quarter left the window, not because a new deal arrived.

    sources123456+3 more
  5. 2026-09-11concentrating to unclearconf 55 → 55 · Claude (model)

    Recounted on 11 Sep 2026 with only buyers in the model or compute layer, as the indicator's definition says (the ledger keeps every deal): the trailing-four-quarter count reads 5, 6, 9, 9 and 9 at the last five quarter ends, the last still open. Two flat steps leave the direction rule reading unclear rather than concentrating.

    sources123456+3 more
  6. 2026-09-10unmeasured to concentratingconf — → 55 · the evaluator (automatic)

    Evaluator: 10 (CNBC, 2026-09-30) reads concentrating over 4 periods with a dead band of 1. Auto-reason; rule rationale: Four quarters; one event is not a trend. More deals per year concentrate the application layer into the labs.

    sources123456+5 more
Why this confidence

Rests on 75 observations from AMD blog, AWS Contact Center Blog, Anthropic newsroom, BetaKit, Business Insider, CNBC, Calcalist Ctech, Cirrus Labs website, CoreWeave newsroom, Crusoe newsroom, Deepgram announcements, Emmi AI news, Fortune, Git AI blog, Google blog (The Keyword), Gretel blog, HKEXnews (listed-company filings), Humanloop website, Jams website, Koyeb blog, Limitless website, Manus blog, Mendral blog, Midjourney announcements, Nebius newsroom, NextSlide website, Nicolas Bustamante's blog (Fintool founder), Nscale press releases, OpenAI research posts, Poolside blog, Rebellions announcements, SEC EDGAR filings (8-K, 6-K, 10-Q exhibits), Silicon Republic, Stilla blog, Stripe newsroom, Suno blog, TechCrunch, The Media Leader FR, The OpenAI Deployment Company newsroom, Weights website, Workshop Labs blog, World Labs blog, d-Matrix announcements; best evidence tier 4, grade A; current.

Confidence is independent of status: 90–95 multiple strong independent sources; 70–89 good evidence, some ambiguity; 50–69 mixed or hard to operationalise; 0–49 limited or vague evidence.

Predictions that lean on this