Scoring rules
Methodology
vc.fail answers a narrow question: of the companies a crypto VC has publicly backed, how much still depends on classical elliptic-curve signatures? The score is a measurement of that public book, not a judgment of the partners, and not a forecast of Q-Day.
The sample uses publicly disclosed investments and official protocol documentation. Missing public evidence is not proof of insecurity, and no score describes a fund's private book or future returns. Read the methodology.
What “exposure” means
Every classified portfolio company gets a posture and a weight between 0 and 100. The fund’s exposure score is the average of those weights across classified deals. If a deal has a disclosed dollar size, capital-weighted scoring is used instead.
| Category | Weight | Rule |
|---|---|---|
| Classical signatures | 100% | Production signatures are classical ECC only (ECDSA, EdDSA, secp256k1, Ed25519, and similar). |
| Hybrid migration | 40–70% | Public PQ migration is underway through hybrid signatures, live PQ components, or a dated roadmap. ECC still remains in the critical path. |
| Post-quantum native | 0–10% | Built from genesis on lattice, hash-based, or other NIST-standardized (or equivalent) post-quantum signatures. |
| Unclassified | Excluded | Not classified yet. Excluded from the diagnostic denominator. |
Primary metric
Exposure score = (sum of exposed capital or deal weights) / (classified portfolio). Unknown names never enter the denominator. That keeps “we haven’t tagged it yet” from looking like safety.
Secondary metrics
- Estimated capital at risk: when a fund has a public AUM figure, AUM × exposure score. This is a proxy, not deployed cost basis.
- Recent still-exposed rate: share of classified deals announced in the last 24 months that are still fully exposed.
- PQ signal: count of hybrid or post-quantum native names. This is the rare positive tell.
Data sources
This first cut is manual. It starts from well-known public investments and protocol documentation, not from a scraped dump of every round. AUM figures are labeled as public estimates (data as of 2026-08-24) and should be treated as order-of-magnitude, not NAV.
- Official fund sites and announcement posts
- Protocol docs for signature schemes
- Public AUM round-ups such as Crypto Fund Research (2026)
- Later: RootData, CryptoRank, Crunchbase / PitchBook for coverage expansion
Limitations
- Public portfolios only. Side letters, undisclosed SPVs, and liquid books are out of scope.
- Completeness is biased toward famous names. A missing PQ-native seed check will overstate exposure until someone sends the source.
- AUM is not the same as capital deployed into the listed companies.
- Hybrid weights are judgment calls inside the 40–70 band. The tag and the source links matter more than the exact integer.
- No private-portfolio claims. If it is not public, it is not here.
Update cadence
Quarterly full refresh, plus ad-hoc updates when a major round or a protocol’s posture changes. Scores recalculate from the JSON in this repo.