Skip to content
Slow Variables

Ask the data

Answers come from the same store as the site; every number is checked against the record it cites.

Indicators

One object, two addresses. Unpublished rows have a definition and data but do not yet meet the publishing rules (a status, counterevidence, and either two sources or one primary source), so they carry no status and no page.

IndicatorBucketLayerStatusConfGradeLatestObs
Rent concentration in semiconductorsCompute & physicalconcentrating85grade A75.0%as of 2026-07-26(2 obs)128
Cloud backlog (remaining performance obligations)Compute & physicalconcentrating80grade A$638Bas of 2026-05-3173
Frontier-lab revenue run-ratesModelconcentrating55grade C$40.0Bas of 2026-08-1334
Lab recoupment ratioModelemerging50grade C0.22×as of 2026-08-13(12 obs)27
Consumer surplus from generative AI (willingness to accept)Labour & consumersdispersing65grade C$172Bas of 2026-03-314
Application layer's share of enterprise generative-AI spend (Menlo)Deployment & applicationemerging55grade B51.4%as of 2025-12-31(2 obs)4
Semiconductors' share of stack operating incomeCompute & physicalconcentrating70grade A57.3%as of 2026-06-30(8 obs)42
Cloud segment operating marginsCompute & physicalstable75grade A39.4%as of 2026-06-30(2 obs)20
Circular financing scaleCompute & physicalconcentrating75grade A$903Bas of 2026-08-17(14 obs)14
Concentration among covered chip filers (HHI)Compute & physicalunclear60grade A0.85 indexas of 2026-06-30(2 obs)13
Concentration among covered cloud filers (HHI)Compute & physicaldispersing60grade A0.35 indexas of 2026-03-31(3 obs)18
Concentration of business AI spend across labs (Ramp, HHI)unpublished
Not yet measurable: the direction rule needs four monthly readings of Ramp's per-lab paid shares and only July 2026 exists; each month arrives as a manual row from Ramp's letter.
Modelno status until publishedgrade B0.46 indexas of 2026-07-31(3 obs)3
Public cloud software valuation multiples (Clouded Judgement)Deployment & applicationemerging50grade B4.40×as of 2026-09-0440
Price per unit of capability (cheapest METR-hour on OpenRouter)Modelnot yet measurable40grade A$0.30as of 2026-09-10(32 obs)32
Frontier labs' share of tracked venture dollarsModelstable50grade A94.8%as of 2026-06-30(32 obs)102
Vertical versus horizontal application fundingDeployment & applicationdispersing45grade A2.30×as of 2026-03-31(11 obs)19
Frontier lab valuation to run-rate multipleModelemerging45grade C28.3×as of 2026-06-16(10 obs)48
Neolab capital raised per dollar of revenueunpublished
Not measurable from free sources: the neolabs' rounds are filed or reported, but none of them has a revenue figure in Epoch or a filing, so the denominator does not exist.
Modelno status until publishedunmeasured0
Lab vertical integration (acquisitions and acquihires, trailing four quarters)Modelconcentrating55grade A10as of 2026-09-30(11 obs)16
Venture flow to revenue lagunpublished
Not yet measurable: the estimate needs at least eight quarters of both venture dollars and revenue for the same sub-layer; the venture series starts in 2023 and revenue exists only for the frontier labs.
Modelno status until publishedunmeasured0
Open-weights gap to the closed frontier (ECI points)Modeldispersing55grade A5.77 indexas of 2026-08-20(3 obs)141
Non-Nvidia share of AI compute shipped (H100-equivalents)Compute & physicalemerging50grade B26.0%as of 2026-03-31(3 obs)51
Hyperscaler capex to AI revenue (trailing year)Compute & physicaldispersing40grade A2.51×as of 2026-06-30(16 obs)48
Hyperscaler server useful lives (10-K estimates)Compute & physicalnot yet measurable65grade A5.00 yearsas of 2025-12-314
Neocloud borrowing cost (CoreWeave facility rates)Compute & physicalnot yet measurable60grade A10.5%as of 2026-06-30(4 obs)12
Enterprise multi-model usage share (Menlo)Modelnot yet measurable45grade B60.0%as of 2023-11-103
Startups' share of enterprise AI application revenueDeployment & applicationemerging50grade B63.0%as of 2025-12-312
Frontier lab gross marginunpublished
Not measurable from free sources: lab gross margins are reported only by The Information and Sacra behind paywalls; the paid stubs exist and are off.
Modelno status until publishedunmeasured0
Compute share of lab spendingunpublished
Not measurable as a series: Epoch published a one-off estimate range, not a table; a manual row can carry the point estimate once quoted.
Modelno status until publishedunmeasured0
Application gross margin net of inference (model tax)unpublished
Not measurable from free sources: needs SemiAnalysis unit inference costs (paid) joined to reported ARR and Ramp token spend; the paid stub is off.
Deployment & applicationno status until publishedunmeasured0
GPU rental price per hourunpublished
Not measurable from free sources: the SemiAnalysis index is paywalled; the paid stub is off.
Compute & physicalno status until publishedunmeasured0
Secondary-market marks between roundsunpublished
Not measurable from free sources: secondary marks are behind Caplight, Forge and EquityZen paywalls; the paid stub is off.
Modelno status until publishedunmeasured0
Revenue per employee in AI-exposed sectorsunpublished
Not measurable yet: no free source pairs sector revenue with employee counts at a usable cadence; BLS industry mappings to exposure are unverified.
Adoptersno status until publishedunmeasured0
Model-layer token concentration on OpenRouter (HHI, top table)Modelnot yet measurable40grade B0.11 indexas of 2026-09-09(19 obs)19
METR 50% time horizonMethodsModelfaster than normal80grade A17.4 h(8.5 h55.1 h)disputedas of 2026-04-0730
METR 80% time horizonMethodsModelfaster than normal70grade A3.1 h(1.6 h6.6 h)as of 2026-04-0726
50%/80% horizon ratioMethodsModelconsistent with normal70grade A6.34×as of 2026-03-05(2 obs)50
US labour productivity, year on yearAdaptationLabour & consumersconsistent with normal80grade A2.2%as of 2026-06-3076
US total factor productivity, private nonfarm businessAdaptationLabour & consumersconsistent with normal75grade A0.8%as of 2025-12-31(2 obs)9
Developer productivity uplift (METR RCTs)ProductsDeployment & applicationconsistent with normal70grade A-4.0%disputedas of 2026-02-242
Enterprise pilots with measurable P&L impactProductsDeployment & applicationconsistent with normal55grade C5.0%disputedas of 2025-06-302
Share of work hours assisted by generative AIEarly adoptionAdoptersconsistent with normal80grade B6.3%as of 2026-06-3027
US firms using AI (Census BTOS)Early adoptionAdoptersconsistent with normal70grade A22.4%(21.7%23.1%)as of 2026-08-0921
US businesses paying for AI (Ramp AI Index)Early adoptionAdoptersemerging60grade B56.1%as of 2026-08-312
Agent-workdays per human workday in frontier research (self-reported)Return arrowModelemerging30grade D3.10×as of 2026-08-151
Human interventions on 4–8 hour agent tasks (self-reported)Return arrowModelemerging30grade D50.0%as of 2026-07-311
Safety brakes inside the labs (events, trailing year)AdaptationModelemerging40grade D3as of 2026-09-01(3 obs)3
Frontier training compute growth (doubling time)MethodsModelemerging60grade B170 daysas of 2025-07-09(66 obs)66
Training power draw growth (doubling time)MethodsCompute & physicalemerging50grade B506 daysas of 2026-08-07(213 obs)213
Hardware price-performance growth (doubling time)MethodsCompute & physicalconsistent with normal40grade B1282 daysas of 2025-11-06(20 obs)20
Inference price at fixed capability (halving time)MethodsModelemerging45grade B68 daysas of 2024-12-13(6 obs)6
Public text data exhaustion year (Epoch projection)MethodsTraining inputemerging40grade B2028 yearas of 2024-06-061
ARC-AGI-2 best scoreMethodsModelemerging55grade B95.0%as of 2026-09-03(203 obs)203
Continual-learning ladder (highest production rung)Return arrowTraining inputemerging45grade D3.00 rungas of 2026-05-27(2 obs)2
FDA-authorised AI-enabled devices, growthAdaptationAdoptersconsistent with normal70grade A26.7%as of 2026-06-30(2 obs)102
US state AI bills introduced (legislative year)AdaptationAdoptersemerging55grade B1035as of 2025-12-312
Active RL-environment vendors (directory count)Return arrowTraining inputemerging35grade B24as of 2026-07-151
Ord half-life misfit (observed 50/80 ratio vs constant hazard)MethodsModelconsistent with normal60grade A2.04×as of 2026-03-05(2 obs)50
FDA-authorised devices built on a large language modelAdaptationDeployment & applicationconsistent with normal60grade A1as of 2025-12-23(3 obs)3
Reliability-science framework adoptionunpublished
Not measurable as a number: adoption is a yes/no per system card, recorded as evidence rows rather than a series.
ProductsModelno status until publishedunmeasured0
Agent calibrationunpublished
Not measurable yet: no public leaderboard reports calibration for agents; HAL and tau-bench report success, not confidence.
ProductsModelno status until publishedunmeasured0
Agent operational safety incidentsunpublished
Not measurable yet: incidents are anecdotal; no incident registry with a denominator exists.
ProductsModelno status until publishedunmeasured0
Deployment data share of trainingunpublished
Not measurable: no lab discloses the provenance mix of its training data; recorded only if a system card states it.
Return arrowModelno status until publishedunmeasured0
Employment in AI-exposed occupations, year on yearAdaptationLabour & consumersconsistent with normal80grade C-0.2%as of 2026-06-303
Entry-level employment shortfall in AI-exposed occupationsAdaptationLabour & consumersemerging70grade B19.0%as of 2026-06-304
Exposed vs unexposed occupation employment gap (Revelio)AdaptationLabour & consumersemerging65grade B-6.0%as of 2026-07-311
California unemployment claims from AI-exposed occupationsAdaptationLabour & consumersconsistent with normal70grade A1.0%as of 2026-07-312
Recent-graduate unemployment rate (NY Fed)AdaptationLabour & consumersconsistent with normal75grade A5.6%as of 2026-06-301
Cross-tracker concordance (labour)AdaptationLabour & consumersconsistent with normal70grade B0as of 2026-07-31(4 obs)4
Augmentation vs automation share (Anthropic Economic Index)Early adoptionAdoptersconsistent with normal60grade B51.4%as of 2026-05-3111
US labour share of income, year on yearAdaptationLabour & consumersemerging65grade A-3.4%as of 2026-06-30(2 obs)38
Customer-support productivity uplift from a generative-AI assistantAdaptationAdoptersconsistent with normal75grade B14.0%as of 2023-04-302
Executives reporting no AI impact on their own firmAdaptationAdoptersconsistent with normal70grade C90.0%as of 2026-02-282
Frontier models completing expert legal tasks end to end (Harvey LAB)ProductsDeployment & applicationemerging40grade D7.1%as of 2026-05-262
Agent reliability across repeated runs (tau-bench pass^k)ProductsModelemerging55grade B56.2%as of 2026-05-184
Progress against the AI 2027 quantitative predictionsMethodsModelemerging50grade B75.0%as of 2026-07-313
Work-related share of ChatGPT consumer messagesEarly adoptionAdoptersemerging65grade B27.0%as of 2025-06-302
Usage as a share of theoretical exposureEarly adoptionAdoptersconsistent with normal60grade B0.49×as of 2026-06-30(2 obs)9
Workers highly exposed to AI with low adaptive capacityAdaptationLabour & consumersemerging45grade B3,300,000 workersas of 2026-01-311
Occupational-mix dissimilarity since ChatGPT (Yale Budget Lab)AdaptationLabour & consumersconsistent with normal65grade B5.26 ppas of 2026-03-01103
Chatbot effect on earnings and hours in exposed occupations (Denmark)AdaptationLabour & consumersconsistent with normal80grade B0.0%as of 2024-12-313
Freelance writing earnings after ChatGPT (Upwork)AdaptationLabour & consumersconsistent with normal70grade B-5.2%as of 2023-12-013
Wage premium for AI skills (PwC AI Jobs Barometer)AdaptationLabour & consumersemerging50grade C62.0%as of 2025-12-312
Expert-data market run-rate (Mercor)MethodsTraining inputfaster than normal70grade C$2.0Bas of 2026-06-302
Internal-to-public deployment gap at the frontierProductsModelconsistent with normal45grade C4.00 monthsas of 2026-06-302
Waymo paid rides per weekAdaptationDeployment & applicationconsistent with normal65grade C500,000 rides per weekas of 2026-03-272