Skip to content
CohortWatch

AssemblyAI

assemblyai.com
YC · S17AI & ML InfrastructureAI-native

65 people

Updated Jul 14, 2026

The reading

59out of 100
HoldingMedium confidence

vs 289 peers · YC · AI & ML Infrastructure · older vintages pooled

includes code signals as of Jun 17, 2026

A level reading. The file carries no history yet, so no trend is claimed.

The published read

Holding steady against its YC AI & ML Infrastructure peers.

Cross-checks agree: all 6 checks run agree. 6 checks · details

Independent reads of this company were compared. Each check below names what was compared and what it found.

  • agreeAI claim vs code evidence. AI claim is corroborated by code evidence. · cohortwatch:ai_scores
  • agreeAI level vs ML hiring/build motion. AI-native level is backed by ML hiring/build motion. · cohortwatch:ai_scores · observations
  • agreeProprietary ML vs third-party models (L2/L3 line). L3 label is backed by proprietary-ML signals. · cohortwatch:ai_scores
  • agreeRetrofit ceiling (founding era). AI level not capped by the retrofit ceiling. · cohortwatch:companies · cohortwatch:ai_scores
  • agreeSingle-source AI signal (attribution guard). AI level rests on independent signals from more than one evidence channel. · cohortwatch:ai_scores
  • agreeBuild depth but quiet momentum. Build depth and momentum are consistent. · cohortwatch:scores · cohortwatch:ai_scores
cohort position not stated: the older-vintage peer pool is too coarse for a fair rank
0100

AssemblyAI is compared within 289 peers (YC · AI & ML Infrastructure · older vintages pooled): cohort position not stated: the older-vintage peer pool is too coarse for a fair rank.

Why this score

Holding steady against its YC AI & ML Infrastructure peers. Strongest signal: code (85/100). Capital collection is inactive in this export, so no capital observation affects the displayed score.

Weights: Hiring 40% · Code 30% · Capital 30%. Updated Jul 14, 2026 · planned refresh ~Jul 21 · score version momentum-v2.

Signal record

Signal breakdown

Hiring 40% · Code 30% · Capital 30%, renormalised over the families present in the scoring record. Each meter is that family’s 0-100 sub-score. Direct source links appear when the public export carries them.

Hiring
weight 0.40
Hiring signal present; 64/100 vs cohort peersGreenhouse
64/100
Code
weight 0.30
Code & package adoption present; 85/100 vs cohort peersGitHubHugging FacenpmPyPI
85/100
Capital
weight 0.30 · one-sided
Collector inactive in this export: no capital observation; no effect on displayed score.
N/A

Linked evidence

Available public sources

This ledger shows the public source links carried by the export.

History

Momentum over time

readings on file: 00–100 scalethe ledger

Not enough history to measure a trend yet tracked weekly across active sources.

0 of 6 weekly readings. A trend appears at 6

File a correction →How is this calculated?