Scale AI (SEAL)
Scorecard
Builds and runs evaluations for CAISI and other governments; Meta acquired 49% in 2025.
VC-backed, San Francisco, US. Vendor / red-team co.; listed with commercial and first-party. Confidence high. Domains: benchmarks, cyber, bio, misuse.
8/8 evidenced · Retrieved & confirmed Scores by preset: Lab procurement –; Regulator or auditor –; Public trust –; Equal weights –. Weakest evidenced dimension: funding 0/4. At least one evidenced conflict dimension at 0: not a candidate for an independence-critical role until the floor moves. Access and methods never set a band.
The weighted numbers are hidden until you ask for them: two readers with different priorities see different numbers, and neither is the site's.
What would move the score. Ownership separation of the evaluation unit; until then the ceiling is low.
- Funding 0 to 1: score –
- Governance 0 to 1: score –
- Personnel 0 to 1: score –
Dissent, lower. Personnel (1 to 0). The founder joined Meta as an executive under the $14.3B ownership transaction and, by Scale's own announcement, keeps his seat on Scale's board. No recusal is documented anywhere in the public record. P.7 routes leadership that moved to a lab under an ownership transaction to P.1, which gives 0 without a recusal; P.4 separately caps a lab employee on the evaluator's board at 2.
Dissent, higher. Funding (0 to 1). Meta's 49% stake is non-voting, Scale's own announcement calls it a minority of outstanding equity, and the interim CEO stated Scale 'remains, unequivocally, an independent company'. If the 20% threshold in F.1 is read as requiring control rather than economic interest, the record supports anchor 1: a material investment from a frontier developer without voting power.
Each value is the tightest admissible cap or, with no cap, the highest admissible floor, under RULES.md. The binding signal is highlighted. A dash means no admissible signal sets a bound under the Retrieved & confirmed policy.
- Nearly half owned by Meta, itself a frontier developer. caps at 0 by F.1binding“Meta has a 49% stake in Scale after its $14.3 billion investment”
Third-Party Model Auditing (Longterm Wiki, 2026-01-29) tier 4 press (aggregator)confirmed Why the U.S. Needs an Independent AI Evaluation Framework for National Security (Scale AI, 2026-05-20) tier 3 self (scale)confirmed Scale AI not winding down following Meta deal, interim CEO says (CNBC, 2025-06-18) tier 4 press (primary)confirmed Scale AI Announces Next Phase of Company’s Evolution (Scale AI, 2025-06-12) tier 3 self (scale)confirmed2026-09-14 - Meta paid $14.3B for a 49% non-voting stake in June 2025; the founder moved to Meta. caps at 0 by F.1“Meta has a 49% stake in Scale after its $14.3 billion investment”
evaluators.csv (24 evaluators: Coefficient and SFF totals, lab money, government contracts, leadership) (Kevin Bass (GitHub), 2026-09-14) tier 2 ledgerimported Scale AI not winding down following Meta deal, interim CEO says (CNBC, 2025-06-18) tier 4 press (primary)confirmed2026-09-11
- Primary business is selling data and evaluation services to the labs; no public COI policy addressing the Meta relationship. caps at 1 by G.2“Scale has worked with the world's leading AI model builders”
Why the U.S. Needs an Independent AI Evaluation Framework for National Security (Scale AI, 2026-05-20) tier 3 self (scale)confirmed2026-09-14 - Meta holds a 49% stake after a $14.3B investment; the founder left for Meta and an interim CEO was named; Google, OpenAI and xAI reportedly reduced engagements. caps at 0 by G.6binding“Meta has a 49% stake in Scale AI after investing $14.3 billion”
Scale AI not winding down following Meta deal, interim CEO says (CNBC, 2025-06-18) tier 4 press (primary)confirmed2025-06-18
- Leadership moved to Meta with the 2025 investment. caps at 0 by P.7 (via P.1)binding“a deal that included the departure of Scale AI founder Alexandr Wang for the social media company”
Third-Party Model Auditing (Longterm Wiki, 2026-01-29) tier 4 press (aggregator)confirmed Scale AI not winding down following Meta deal, interim CEO says (CNBC, 2025-06-18) tier 4 press (primary)confirmed Scale AI Announces Next Phase of Company’s Evolution (Scale AI, 2025-06-12) tier 3 self (scale)confirmed2026-09-14 - After joining Meta, the founder continues to serve as a director on Scale's board, per Scale's own announcement of the investment. caps at 2 by P.4“Wang will continue to serve as a director on the Scale Board of Directors”
Scale AI Announces Next Phase of Company’s Evolution (Scale AI, 2025-06-12) tier 3 self (scale)confirmed2026-09-15 - Document request: Request from Scale AI the board-recusal terms, if any, that apply to Alexandr Wang's continuing directorship while he is employed by Meta (Scale announcement, 12 June 2025), and whether he sits on any committee reviewing evaluation work. Per RULES 12 this goes to the person before publication.
- Government-defined risk domains and access agreements; first private evaluator authorized by the US institute. floors at 3 by A.4binding“Scale AI SEAL : Safety, Evaluation, and Alignment Lab; first third-party evaluator authorized by US AISI”
Third-Party Model Auditing (Longterm Wiki, 2026-01-29) tier 4 press (aggregator)confirmed Why the U.S. Needs an Independent AI Evaluation Framework for National Security (Scale AI, 2026-05-20) tier 3 self (scale)confirmed2026-09-14
- Scope set by government or lab clients. caps at 2 by S.2binding“The government defines the risk domains it wants evaluated”
Why the U.S. Needs an Independent AI Evaluation Framework for National Security (Scale AI, 2026-05-20) tier 3 self (scale)confirmed2026-09-14
- Results for governments not public. caps at 2 by R.4 (by analogy)binding“The government provides subject matter expertise and interprets the results for policy and standards”
Why the U.S. Needs an Independent AI Evaluation Framework for National Security (Scale AI, 2026-05-20) tier 3 self (scale)confirmed2026-09-14
- Benchmarks partly public (SEAL leaderboards); government evaluations closed. caps at 2 by M.2binding“Scale Labs builds the benchmarks and conducts the technical testing”
Why the U.S. Needs an Independent AI Evaluation Framework for National Security (Scale AI, 2026-05-20) tier 3 self (scale)confirmed2026-09-14
- Sells data and evaluation services to labs. caps at 1 by X.2binding“Meta has been one of Scale AI's biggest customers”
Why the U.S. Needs an Independent AI Evaluation Framework for National Security (Scale AI, 2026-05-20) tier 3 self (scale)confirmed Scale AI not winding down following Meta deal, interim CEO says (CNBC, 2025-06-18) tier 4 press (primary)confirmed2026-09-14
Values under each evidence policy
| Dimension | Leads included | Retrieved & confirmed | Against interest | Verified spans | Primary only |
|---|---|---|---|---|---|
| Funding | 0 | 0 | 0 | 0 | – |
| Governance | 0 | 0 | 0 | 0 | – |
| Personnel | 0 | 0 | 0 | 0 | – |
| Access depth (lab-granted) | 3 | 3 | 3 | 3 | – |
| Scope control | 2 | 2 | 2 | 2 | – |
| Publication rights | 2 | 2 | 2 | 2 | – |
| Method transparency | 2 | 2 | 2 | 2 | – |
| Role incompatibility | 1 | 1 | 1 | 1 | – |
| Score, Lab procurement | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Unevidenced, 0/8 |
| Score, Regulator or auditor | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Unevidenced, 0/8 |
| Score, Public trust | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Unevidenced, 0/8 |
| Score, Equal weights | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Disqualifying floor, 8/8 | – Unevidenced, 0/8 |
Traced money and ties
hop 0: 1 row (investment $14300.0M). Confirmed rows: 1 of 1 Dollar sums: confirmed + unaudited USD rows only; imported figures are not re-derived and never summed. Second hop traced for 1 of 1 sources. Ties: Meta (investor, distance 0, imported). Checked and not found: 1.
Ledger
Money in
- T52 from Meta: investment $14.30B, 2025-06. 49% non-voting stake confirmedtier 4 presshttps://www.cnbc.com/2025/06/18/scale-ai-not-winding-down-following-meta-deal-interim-ceo-says.html
Checked and not found
- N12 Does Coefficient Giving fund Scale AI?
Searched Coefficient grants index, 2,911 rows, snapshot 2026-09-11: no grant from Coefficient Giving in its index. Prompted by imported lead: kevinnbass ledger row EV16. importedhttps://coefficientgiving.org/grants/
Roles hosted
- R51 Meta: investor, 2025-06. 49% non-voting; imported kevinnbass/metr-money-figure:EV16 importedhttps://scale.com/
- R92 Jason Droege: principal, 2025-06. research pass 2026-09-15; interim CEO unauditedhttps://www.cnbc.com/2025/06/18/scale-ai-not-winding-down-following-meta-deal-interim-ceo-says.html
One hop away
Funding graph, focused
Every one-way chain to a lab and to an evaluator at full strength, everything else faded. Hover any name to trace its connections; click a name to open its page.