"A ranking is only as credible as its willingness to say what it can't see."
Most AI rankings are editorial lists dressed up as data. Bluescore is built the other way around: it starts from observable signals, turns them into six dimension scores, and ranks every company on the same scale. There's no hidden factor and no machine-learning model deciding the rank. This post explains methodology v4.0, which took effect on October 18, 2026, and replaces our earlier description of the framework.
Step 1: Collect the Signals
Eight research agents and our data feeds record each company's raw signals: funding, hiring, technical output, news, partnerships, revenue, leadership, and competitive position. Funding totals are written only when a cited source names the company, and each news match is confirmed before it counts.
Step 2: Clean Them Up
For each signal we take the latest known value within its lookback window: 14 days for most signals, 21 for the weekly funding fields, and 45 for total funding. Signals estimated by AI are smoothed with a trimmed mean (drop the highest and lowest readings, average the rest), so a rank moves when the company changes, not when a single day's reading wobbles. Whole-company figures such as revenue, patents and papers are attributed to the AI share: AI Native 100%, Applied AI 60%, AI Adopter 25%, Big Tech 20%.
Step 3: Six Dimensions
Each signal is converted to a z-score across all 250 companies (log-scaled for counts) and mapped to 0–10 through the normal distribution, so a company at the population average scores 5. A missing signal takes the population median, so gaps neither help nor hurt. The signals then roll up into six dimensions, each with a fixed weight:
D1 – Business Model & Revenue (21%). ARR signal, paying customers, and pricing and enterprise tiers.
D2 – Product Depth (21%). Model and product releases, open-source models, and API integrations.
D3 – Data & Model Moat (19%). Patents, research papers, and model releases.
D4 – Talent & Research (16%). Headcount, engineering share of roles, and GitHub activity.
D5 – Customer Adoption (13%). News reach, G2 reviews and rating, web traffic, and market share.
D6 – Investment Momentum (10%). Total funding, the last round, and 6-month stock change for public companies.
Model releases are counted from a release ledger: every named release from the last 12 months, with a release seen only once dropping out after 7 days.
Step 4: One Score, One Ranking
The Bluescore is 10 times the weighted sum of the six dimension scores, a number from 0 to 100. Every company is ranked against all 250 at once, so a 70 means the same thing in every part of the index.
Each company also sits in one of four scoring cohorts: AI Native, Applied AI, AI Adopter and Big Tech, which give context rather than changing the score:
AI Native: pure-play AI companies founded to build AI products and infrastructure. Big Tech: incumbents with major AI divisions. Applied AI: AI applied to physical or specialized domains like robotics, defense, biotech, and industrial automation. AI Adopter: established software firms integrating AI into existing products.
The rankings show each company's rank within its cohort next to its overall rank. Both come from the same Bluescore.
Step 5: Entity Status
A score is meaningless if the company no longer exists as an independent entity. Acquired, merged, taken-private, or shut-down companies are removed from the active index, which led to more than 40 removals in September alone. Acquisitions announced but not yet closed are flagged for review.
What Bluescore Doesn't Measure Yet
A seventh dimension, Responsible AI, is defined but not yet scored. Until a reliable source exists, its weight is shared across the six scored dimensions rather than filled with a placeholder.
What Bluescore Is Not
Bluescore is relative, not absolute: it reflects a company's position among the AI250. It's not a valuation model. Several signals are estimated by research agents rather than read from filings, and companies with thin public footprints score near the middle on many of them. A genuine change can take one to two weeks to show fully in rank. And it's not a recommendation to buy, partner, or procure. It's a structured signal designed to inform research, not replace due diligence.
Cadence and Governance
Scores and ranks are recomputed every night. Signals refresh on a schedule matched to how fast they change: news daily, funding weekly, and the rest every two or three days. Weights change only through a reviewed, versioned methodology update with an effective date. Every company profile includes a "Why this Bluescore?" breakdown showing how much each dimension contributes to its score.
Read the full methodology page for the formula, the signals behind each dimension, and the theoretical foundations.