How the Ledger Score is calculated.

The four dimensions

Every meaningful quantum company is scored on four dimensions, each 0–100. The four scores are weighted into a composite 0–100 total.

• Technology (30%) — Is the underlying physics actually working at scale, and is the architecture credible to FTQC?
• Capital (20%) — Can they fund the next milestone without diluting their option value?
• Commercial (30%) — Is revenue real, growing, and diversified — or is it narrative?
• Government validation (20%) — Has the most rigorous independent technical buyer (DARPA, NIST, DOE, equivalents abroad) blessed the approach?

Weights — why these and not others

Tech and Commercial are equal-weight at 30% each. Quantum is a deep-tech sector where engineering progress and revenue are independent signals on different time-scales — overweighting either alone produces a distorted ranking. A company can have superb tech and no commercial traction (PsiQuantum) or modest tech and growing revenue (D-Wave annealing); both deserve credit for what they have, penalty for what they lack.

Government validation gets 20% — large for a non-financial input. DARPA QBI is the only public program with an independent IV&V team auditing every performer's roadmap. The signal is uniquely costly to fake. We weight it heavily but not above commercial reality.

Capital is 20% — high enough to matter, low enough to avoid double-counting valuation. The score is not a valuation; it is a quality assessment. A great company with a depleted balance sheet is still a great company — the capital score penalizes runway risk, not stock-price gravity.

Dimension 1 — Technology score (0–100)

The tech score is built from three component subscores that average to the dimension total.

1. Best published 2-qubit gate fidelity (40% of tech). Linear from 99.0% (= 0 points) to 99.99% (= 100 points). Below 99% scores zero. Source: company press releases, peer-reviewed papers, DARPA disclosures.
2. Logical qubits demonstrated (35% of tech). Logarithmic curve: 1 logical qubit = 20 points, 10 = 60 points, 48+ = 95 points, 100+ = 100 points. Logical here means error-corrected with active syndrome extraction, not algorithmic placeholder.
3. Coherence/gate-time ratio + architecture credibility (25% of tech). Judgment call based on documented coherence times, native connectivity, scaling path to FTQC. Companies with no public credible FTQC roadmap score zero on this component regardless of physical qubit count.

Annealing platforms (D-Wave) are scored against a different rubric on subscore 2 because logical qubits don't apply. The methodology page is honest about that — annealing tech scores have limited cross-comparability to gate-model tech scores. We score them anyway because they exist in the same investable universe.

Dimension 2 — Capital score (0–100)

Three component subscores:

1. Runway in years (50% of capital). Cash divided by trailing burn. 4+ years = 100 points, 1 year = 30 points, <6 months = 0 points. Parent-company-funded units (IBM Quantum, Microsoft Quantum, Google Quantum AI) score 100 by default because the parent balance sheet is the runway.
2. Recent valuation momentum (30% of capital). Is the company raising at flat, up, or down rounds? Up rounds in the trailing 18 months score full points; flat rounds score 50; down rounds score 0. Public companies use stock-price-vs-IPO as a proxy.
3. Market access (20% of capital). Can the company tap public, private, or government capital efficiently? Public companies with at-the-market facilities score full points. Privates with deep-pocketed strategic backers score high. Companies without obvious next-funding path score low.

Dimension 3 — Commercial score (0–100)

Four components, summed to 100:

1. TTM revenue magnitude (30 points). Logarithmic from $0 (= 0 points) to $200M+ (= 30 points). Adjusted for accounting category — system sales recognize lumpily, subscription revenue gets a small premium for predictability.
2. Revenue growth YoY (30 points). Linear from −20% (= 0 points) to +200% (= 30 points). Declining revenue is penalized; hypergrowth off a tiny base earns the full 30 only when the absolute number is also non-trivial.
3. Named customer base (30 points). Number and quality of disclosed paying customers. Fortune 500 / nation-state customers score higher than university research partnerships. Press-release-only relationships score zero.
4. Concentration penalty (−10 points). Concentration above 40% of revenue from a single customer triggers a flat −10. Above 60%, −20. This is why Quantinuum (RIKEN 60%) loses 10 points despite strong revenue growth.

Dimension 4 — Government validation (0–100)

Four buckets, additive:

1. DARPA QBI / US2QC stage (50 points). US2QC final (Stage C equivalent) = 50; QBI Stage B = 35; QBI Stage A only = 15; QBI Stage A eliminated = 0.
2. NATO / EU / AUKUS programs (20 points). EU Quantum Flagship, UK NQCC, Canadian NRC, AUKUS Quantum Arrangement, etc.
3. Major government contracts (15 points). Named USAF, DOE, NIST, AFRL contracts with disclosed values.
4. Standards body and IV&V participation (15 points). NIST PQC submission, participation in IEEE/ITU quantum standards, ISO/IEC quantum security working groups.

What the score does NOT include

• Stock price. Price is the market's current valuation of the score. The score is the input, not the output. A high-scoring company can be overpriced; a low-scoring company can be cheap. Discussion of price vs score belongs in the per-company writeup, not the score itself.
• Patents. Patent activity is a noisy signal in deep tech — granted patents lag, filings can be defensive, and quality varies wildly. We track patents in the per-company profile but do not include in the score.
• Hype / press volume. Press coverage correlates with PR budget, not technical quality. Excluded on principle.
• Founder pedigree. Important but covered in the per-company qualitative writeup.

Update cadence and change logging

Scores are refreshed on the 13th of each month. Out-of-cycle updates occur when material events land that change a subscore by ≥10 points — DARPA stage decisions, earnings prints, S-1 filings, major technical demonstrations.

Every score change is logged in a changelog at the bottom of each per-company profile. Historical scores are preserved so movements are visible. We do not retroactively edit prior scores; if a past scoring was wrong, we publish a correction note rather than rewriting history.

Honest limitations

• Information asymmetry favors public companies. Private-company commercial and capital subscores rely on the most credible third-party estimates available. We flag every assumption inline. Where a number is unknown, we score conservatively rather than guess.
• Annealing vs gate-model. Cross-comparability of tech scores between modalities is limited. D-Wave's annealing tech score is not directly comparable to IonQ's gate-model tech score. The composite is still useful for ranking commercial viability today; it should not be read as "X has better physics than Y" across modalities.
• Lag on commercial recognition. Lumpy enterprise system sales make TTM revenue noisy quarter-to-quarter. We use trailing 4Q windows where possible to smooth.
• The score is opinion, not advice. The Ledger Score is a structured analytical opinion published for transparency. It is not investment advice. Read the per-company profile, read the source citations, form your own view.

Disagree with a number? Disagree with a weight? Email Connor at the address in the about page. The methodology improves when it's pressure-tested.

Return to the live ranking.