The stack
Seven layers, twenty-four sub-layers, and the companies the tracker follows in each. Membership is many-to-many and dated; an entity is verified once a filing or a dataset row attaches to it. Only the compute layer has segment filings today, so most sub-layers are lists of names waiting for a series.
1. Compute & physical · The chip makers, the big cloud companies and the newer rented-computing firms, data centres and the power they draw, memory and networking.
- 1. Semiconductors & AI chip startups29 entities · 11 verified · 6 indicatorsRent concentration in semiconductors concentratingSemiconductors' share of stack operating income concentratingChips' share of the stack's gross profit concentratingConcentration among covered chip filers (HHI) stableNon-Nvidia share of AI compute shipped (H100-equivalents) emergingHardware price-performance growth (doubling time) consistent with normal
- 2. Hyperscalers27 entities · 11 verified · 7 indicatorsCloud backlog (remaining performance obligations) concentratingCloud segment operating margins concentratingConcentration among covered cloud filers (HHI) stableHyperscaler capex to AI revenue (trailing year) dispersingHyperscaler server useful lives (10-K estimates) emergingNeocloud borrowing cost (CoreWeave facility rates) emergingGPU rental price per hour emerging
- 3. Data-centre developers & power66 entities · 6 verified · 1 indicator
- 4. Memory/HBM25 entities · 5 verified
2. Model · The labs that build AI models: the leaders such as OpenAI and Anthropic, newer labs founded since 2023, and labs that publish their models openly, many of them Chinese.
- 5. Frontier labs7 entities · 5 verified · 12 indicatorsFrontier-lab revenue run-rates concentratingLab recoupment ratio emergingLab concentration of routed tokens (OpenRouter, HHI) emergingPrice per unit of capability (cheapest METR-hour on OpenRouter) emergingPrice per unit of capability (cheapest near-frontier model, Artificial Analysis) emergingFrontier labs' share of tracked venture dollars stableFrontier lab valuation to run-rate multiple emergingLab vertical integration (acquisitions and acquihires, trailing four quarters) emergingEnterprise multi-model usage share (Menlo) emergingFrontier lab gross margin (estimated) emergingModel-layer token concentration on OpenRouter (HHI, top table) emergingInternal-to-public deployment gap at the frontier consistent with normalAnthropicGoogle DeepMindMeta PlatformsOpenAIxAIAlphabet Inc.Scale AI
- 6. Neolabs (founded 2023+)131 entities · 13 verified1XAaruAdaption LabsAdvanced Machine Intelligence / AMI LabsAI21 LabsAleph AlphaAntim LabsApptronikall companies
- 7. Open-weight & Chinese labs8 entities · 4 verified · 2 indicatorsOpen-weights gap to the closed frontier (ECI points) unclearOpen-weights lag behind the closed frontier (months) emerging01.AIAlibaba QwenBaichuanDeepSeekMiniMaxMoonshot AIStepFunZhipu AI
3. Training input · What models are trained and tested on: work by paid human experts, practice tasks for training by trial and error, generated data, licensed content, and tests.
- 8. Human/expert data & labelling22 entities · 2 verified · 2 indicatorsLabs buying from the training-input layer (enter bell) emergingExpert-data market run-rate (Mercor) emergingAppenBee MapsCloudFactoryDatacurveHandshake AIHuman ArchiveiMeritInvisible Technologiesall companies
- 9. RL environments15 entities · 0 verified · 1 indicatorAntiochApplied ComputeApplied IntuitionBespoke LabsEndlessFleetHalluminateJudgment Labsall companies
- 10. RLaaS / post-training / continual-learning infra11 entities · 2 verified · 3 indicators
- 11. Synthetic data & pipelines9 entities · 0 verified
- 12. Permissioned data markets & licensing12 entities · 1 verifiedCreated by HumansDataradeDecentriqDefined.aiDuality TechnologiesHuman NativeProRataProtegeall companies
- 13. Evals19 entities · 3 verified
4. Serving & orchestration · The businesses that run models for others, route requests between them, and supply the plumbing for AI agents.
- 14. Inference serving43 entities · 4 verified · 1 indicatorInference-layer revenue run-rates emerging
- 15. Routing13 entities · 3 verified
- 16. Agent infrastructure42 entities · 1 verified
- 17. Agent security & identity39 entities · 0 verified
5. Deployment & application · The products people use and the firms that put AI to work: coding assistants, apps built for one industry, consumer apps, AI-native firms that sell finished work, firms that train the people who deploy AI, and firms that buy up service businesses to run them with AI.
- 18. Deployment / FDE / implementation services35 entities · 3 verified · 1 indicatorEnterprise pilots with measurable P&L impact consistent with normalAirbyteAmpersand LabsAnthropic enterprise services ventureArtiumBrain Co.Caritas Venture Co.CiridaeCodosall companies
- 19. Talent production / AI education & FDE training40 entities · 0 verified
- 20. Coding agents & dev tools17 entities · 3 verified · 1 indicatorDeveloper productivity uplift (METR RCTs) consistent with normal
- 21. Vertical AI by domain48 entities · 12 verified · 3 indicators
- 22. Autonomous services / AI-native service firms91 entities · 0 verified
- 23. Consumer AI7 entities · 2 verified · 1 indicatorWaymo paid rides per week consistent with normalCharacter.AIHiggsfieldInstinctKrea AIReplikaSpeakPerplexity
- 24. AI-native rollups / buyout vehicles170 entities · 0 verified
6. Adopters · The companies and people using the products. Tracked through its readings; no company list.
Indicator-only layer: no companies, only measures of adopters, workers and consumers.
7. Labour & consumers · Workers and consumers: wages, jobs, and the value users get without paying for it. Tracked through its readings; no company list.
Indicator-only layer: no companies, only measures of adopters, workers and consumers.