Skip to content
Slow Variables
Ask

Model

Frontier labs' share of tracked venture dollars

Frontier labs' share of all primary-round dollars across the tracked stack over the trailing four quarters, from Form D filings where entities file and Epoch's press-compiled rounds elsewhere. Derived metric frontier_share_of_venture_dollars.

Reason through this

The reading

The denominator is the seed list, so the level is high by construction; the direction says whether the rest of the stack is catching up.

Direction over the last 5 readingseach dot links to its record
in the rule's windowearliermoves within ±5.0 pts read as stable, and a move counts only if more than half the steps go its way; this window moved −0.5 pts and 2 of 4 steps fell; higher reads concentrating
estimate

15 readings drawn from the derived metric frontier_share_of_venture_dollars; 117 observations in the evidence.

Derived (frontier_share_of_venture_dollars) from: Epoch AI data hub — AI companies; SEC EDGAR Form D (private offerings). Epoch AI, 'Data on AI Companies' (CC BY 4.0), epoch.ai/data. SEC EDGAR Form D data sets.

The numbers
as ofreadingvaluerecord
2026-09-30—91.2%estimate
2026-06-30—94.1%estimate
2026-03-31—95.9%estimate
2025-12-31—92.4%estimate
2025-09-30—91.7%estimate
2025-06-30—91.6%estimate
2025-03-31—89.7%estimate
2024-12-31—84.4%estimate
2024-09-30—73.8%estimate
2024-06-30—75.0%estimate
2024-03-31—74.5%estimate
2023-12-31—91.2%estimate
2023-09-30—95.3%estimate
2023-06-30—95.3%estimate
2023-03-31—99.4%estimate

Why it matters

Capital concentrating in the model layer's incumbents is the installation-phase pattern; capital spreading to applications and infrastructure is what a maturing stack looks like (P1 seed indicator 8).

What cuts against it

Frontier rounds come mostly from press (Epoch, tier 5) because the labs raise through vehicles that do not file Form D; many seed companies have no CIK matched yet, so the denominator undercounts.

How it is measured

Four quarters; a move under five points is a single round's timing. Rising = capital concentrating in the frontier labs.

Applied to metric:frontier_share_of_venture_dollars.

Derived rows (15)
as ofdimsvalueinputs
2026-09-3091.2%sources123+38 more
2026-06-3094.1%sources123+36 more
2026-03-3195.9%sources123+37 more
2025-12-3192.4%sources123+31 more
2025-09-3091.7%sources123+32 more
2025-06-3091.6%sources123+29 more
2025-03-3189.7%sources123+25 more
2024-12-3184.4%sources123+26 more
2024-09-3073.8%sources123+23 more
2024-06-3075.0%sources123+22 more
2024-03-3174.5%sources123+25 more
2023-12-3191.2%sources123+20 more
2023-09-3095.3%sources123+12 more
2023-06-3095.3%sources123+11 more
2023-03-3199.4%sources123+1 more
Latest readings
  • 2026-09-30 · 91.2%as of 2026-09-30(41 obs)
  • 2026-06-30 · 94.1%as of 2026-06-30(39 obs)
  • 2026-03-31 · 95.9%as of 2026-03-31(40 obs)
  • 2025-12-31 · 92.4%as of 2025-12-31(34 obs)
  • 2025-09-30 · 91.7%as of 2025-09-30(35 obs)
  • 2025-06-30 · 91.6%as of 2025-06-30(32 obs)
Why this status, and its history

Evaluator: 0.9479 (SEC EDGAR Form D (private offerings), 2026-06-30) reads stable over 4 periods with a dead band of 0.05. Auto-reason; rule rationale: Four quarters; a move under five points is a single round's timing. Rising = capital concentrating in the frontier labs.

  1. 2026-09-10unmeasured to stableconf — → 50 · the evaluator (automatic)

    Evaluator: 0.9479 (SEC EDGAR Form D (private offerings), 2026-06-30) reads stable over 4 periods with a dead band of 0.05. Auto-reason; rule rationale: Four quarters; a move under five points is a single round's timing. Rising = capital concentrating in the frontier labs.

    sources123456+26 more
Why this confidence

Rests on 117 observations from Epoch AI data hub — AI companies, SEC EDGAR Form D (private offerings); best evidence tier 4, grade B; 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