Methodology & Data Integrity

Hiring Momentum Data
Methodology & Quality Standards

We disclose our full collection methodology, signal definitions, and quality-control standards.
Independent verification by institutional investors and quantitative researchers is welcomed.

Ver 1.1 — June 2026  |  ASI Inc. / Physical AI Research Team  |  info@humanoid-jobs.com
430
Companies Tracked
Daily
Collection Frequency
40,481+
Total Snapshots
2019–
Data Inception

Data Sources

All data is collected directly from each company's official applicant-tracking system (ATS) via primary-source polling. We do not aggregate from job boards, LinkedIn, or any secondary source.

Why This Dataset Is Proprietary

This dataset is built on humanoid-jobs.com, a job board dedicated to the Physical AI vertical. Because collection, taxonomy, and signal definitions are purpose-built for this single domain rather than inherited from a generic aggregator, every layer is optimised for the robotics / Physical AI context.

Continuous daily collection compounds into a proprietary time series that a later entrant cannot reconstruct retroactively. What a company was hiring on a given past date is unrecoverable unless it was observed that day. The longer the dataset runs, the scarcer and less replicable its history becomes — time itself is the moat.

PAHI & Commercialization Index Methodology

The PAHI (Physical AI Hiring Index) is a hiring-count index computed on a fixed, like-for-like constituent panel to remove spurious moves caused by changes in the number of tracked companies. It is normalised to the base week (week of 2026-05-19) = 100. Cohort criteria: companies with daily data in every week since the base week and a base-week average of at least 10 open roles (excluding companies under-captured during collection ramp-up). For each week, the cohort total is divided by the base-week total and multiplied by 100. The formula and constituents are fully public.

The Commercialization Index classifies each job title via deterministic keywords into a function axis (research / engineering / manufacturing-production / field-deployment / GTM-sales / corporate). The commercialization score is (manufacturing + field + GTM) / total. A rising ratio is a leading indicator of a company moving from the R&D phase toward production and commercialization. Classification is deterministic (not a black box); role composition is stored daily in js_job_composition and reconstructed historically from job-title history. This signal is producible only because the parent of the dataset is a Physical-AI-specialist board, and cannot be replicated from generic hiring data.

Long-history views (provisional): the weekly PAHI and Commercialization Index begin only at our collection start (2026-05), so their history is shallow. To compensate, we separately provide provisional long-range views built from past job titles in the Wayback Machine. PAHI-Deep (3-year view) is computed on a single, consistent basis from Wayback counts only for a fixed panel of six companies that trace back years (Shield AI, Diligent Robotics, Figure AI, Formic, Dexterity, Bright Machines), normalised to 2023-02 = 100 (monthly; forward-fill capped at 90 days to drop stale carry-forward). The long-range Commercialization history uses only calendar months in which at least three companies are present — genuine cross-sections, never single-company noise. In both cases we never mix measurement bases (Wayback counts vs. daily counts) and we explicitly label these as provisional.

Signal Definitions

Signals are deterministic, rule-based indicators computed from the 14-day percentage change and absolute change in open job count. No machine-learning models or subjective judgment are involved.

SignalSignal ThresholdInterpretation
🔥 STRONG_SURGE≥ +30% and ≥ +20 postingsRapid hiring expansion — frequently observed following a major product launch or funding event
📈 SURGE≥ +20% and ≥ +10 postingsClear acceleration in hiring activity
↗ RISING≥ +10%Gradual upward hiring trend
→ STABLEWithin ±10%Hiring pace maintained within normal variance
↘ DECLINING≤ −10%Contracting hiring pace
🧊 FREEZEZero open postings (current day)Observed cessation of hiring activity
🧊 FREEZE_CONFZero postings sustained ≥ 30 consecutive daysConfirmed hiring freeze — distinguished from transient zero-count observations
The percentage change is calculated relative to the 14-day trailing moving average. For companies with thin job-count baselines (fewer than 5 postings), absolute change takes precedence over the percentage threshold.

Coverage Universe

The dataset covers430+ companies across the Physical AI sector.

Quality Controls

Limitations & Caveats

Explore the Live Dataset

Access real-time hiring signals for the top 10 companies with a free trial.