Stack / Training input / sub-layer 8
Human/expert data & labelling
Indicators
Venture flow
Form D filings
Primary-round equity per calendar quarter: Form D amount sold where the entity files, Epoch's press-compiled rounds where the entity has no filing in the same quarter or within 45 days. Amendments replace their originals; SPVs and secondaries are excluded; Form D debt is listed in the table, not drawn.
Derived (venture_dollars_by_source) from: SEC EDGAR Form D (private offerings). SEC EDGAR Form D data sets.
Entities · 22
| Entity | Role | Since | Verified | Latest observation |
|---|---|---|---|---|
| Appen | primary | — | ○ not yet | none |
| Bee Maps Own site beemaps.com, read 5 Oct 2026: Runs a fleet of contributor dashcams and sells the fresh street-level video and imagery to mapping and physical-AI companies. Formerly known as Hivemapper. Found in the Part 39 search for the map's empty parts and kept by the staff review. | primary | — | ○ not yet | none |
| CloudFactory | primary | — | ○ not yet | none |
| Datacurve | primary | — | ○ not yet | none |
| Handshake AI | primary | — | ○ not yet | $100M · gross_revenue_run_rate_usd.pt · 2025-08-31 |
| Human Archive Own site humanarchive.ai, read 5 Oct 2026: Sells recordings of people doing physical tasks, including video, depth, touch and motion capture, for training robots. Found in the Part 39 search for the map's empty parts and kept by the staff review. | primary | — | ○ not yet | none |
| iMerit | primary | — | ○ not yet | none |
| Invisible Technologies | primary | — | ○ not yet | none |
| Kled Own site kled.ai, read 5 Oct 2026: Runs an app where individuals are paid to upload their own content and licenses the pooled datasets to AI companies. Own site describes it as a Nitrility Inc. company. The data is contributed by consumers rather than captured from their use of a product. Found in the Part 39 search for the map's empty parts and kept by the staff review. | primary | — | ○ not yet | none |
| Labelbox | primary | — | ○ not yet | none |
| MercorCIK 0002103986 Market map: kadan_ai_native_services, read 28 Sep 2026 | primary | — | ● yes | $2.0B · gross_revenue_run_rate_usd.pt · 2026-07-09 |
| Micro1 | primary | — | ○ not yet | none |
| NATIX Network Own site natix.network, read 5 Oct 2026: Collects multi-camera driving footage from contributors' cars and phones and sells it for autonomous-driving and mapping work. Contributors are rewarded with a crypto token. Found in the Part 39 search for the map's empty parts and kept by the staff review. | primary | — | ○ not yet | none |
| Overmind Own site overmindlab.ai, read 5 Oct 2026: Sells a platform that turns an agent's production traces into cleaned datasets and then into fine-tuned models. Early-stage; trace-to-dataset is the first step of a wider training platform. Found in the Part 39 search for the map's empty parts and kept by the staff review. | primary | — | ○ not yet | none |
| Pareto | primary | — | ○ not yet | none |
| ProlificCIK 0001976659 | primary | — | ● yes | $16.6M · amount_sold_usd.pt · 2023-04-25 |
| Sapien | primary | — | ○ not yet | none |
| Scale AI 49% acquired by Meta in June 2025; human-data business continues. | primary | — | ○ not yet | 1.00 score · signal.pt · 2025-06-12 |
| Snorkel | primary | — | ○ not yet | none |
| Surge AI | primary | — | ○ not yet | 1.00 score · signal.pt · 2025-07-03 |
| Toloka | primary | — | ○ not yet | none |
| Turing | primary | — | ○ not yet | 1.00 score · signal.pt · 2025-03-06 |
Seed lists come from the capture brief and are verified against filings one by one. A name here is not a claim about size or funding until an observation attaches.