$15T is allocated blind.
Capital flows into the physical economy at a rate of fifteen trillion dollars a year — debt, infrastructure, insurance, equipment, pension. Banks underwrite balance sheets. Credit bureaus underwrite individuals. Infra funds underwrite assets. Insurance underwrites cohorts. Nobody underwrites the operator who guarantees the cash flow.
Allometry is the interchange standard between operators and capital. Four lender tiers. One vault. Time-in-vault unlocks each tier. Diligence collapses from quarters to days.
Four tiers. One substrate. Time-in-vault unlocks each.
Each tier underwrites a different question. The covenants change. The horizons change. The minimum statistical mass changes. What does not change is the atomic data unit — the customer site, scored for margin quality, contract terms, and forward retention. If the schema is right from day one, the same site-level record aggregates cleanly to operator (T4), portfolio (T2), cohort (T1), and fund (T3).
Banks, asset-based lenders, equipment finance, working capital
Live Oak, BDC, Siena Lending, Pathward/Crestmark, Wells Fargo Capital Finance, RBCx. Backward-looking on equipment residual + personal credit. 6–10 week diligence today. The vault collapses it to days.
Infrastructure debt funds
Brookfield BID IV ($4B+ first close), Generate Capital, Stonepeak, BlackRock/GIP, Macquarie, Nuveen, Allianz, Foresight. Underwrites the asset and the contract — not the operator. The vault is the 5th credit pillar.
Insurance-linked capital
Munich Re, Swiss Re, Hannover Re, Nephila, Lloyd's syndicates, AXA XL, Markel, Parametrix, Descartes Underwriting. Needs statistical mass — distributions, not points. Allometry becomes the actuarial table.
Pension allocators · sovereign wealth · rated paper
CPPIB, CDPQ, OTPP, CalPERS ($60B climate), CalSTRS, APG, NBIM, GIC, ADIA, Mubadala. NRSRO-rated paper requires full track record + statistical mass + standardized data. The category endgame.
$15T flows. Almost none of it underwrites the operator.
Every layer of the capital stack has its own way of pricing risk in the physical economy. None of them prices the operator. The result is a diligence cycle that takes between six weeks (banks) and twenty-four months (pension allocators), driven by data fragmentation, manual verification, and a structural absence of the underwriting primitive that would otherwise let capital flow.
Six systems, no canonical record
An operator's CRM, ERP, FSM, accounting ledger, dispatch system, and contract repository each hold a partial view. None of them share an ontology. Lenders rebuild the model from scratch every cycle.
Field exams + appraisals + ESAs
A typical bank ABL diligence runs 6–10 weeks: collateral audit (2–4 wk), equipment appraisal (2–3 wk), Phase I ESA (3–6 wk), AR-aging-to-GL reconciliation (1–2 wk). The auditor samples; the operator absorbs the cost.
One operator can't fit a distribution
Insurance requires loss-distribution shape, not point estimates. Pension paper needs 100+ comparable obligors with 36+ months realized data. Until the cohort exists, the upper tiers are mathematically locked.
Trust without verification
Quarterly servicer certificates. Annual loss runs. Borrower-prepared backlog schedules. Lenders price the trust gap — 25–75 bps of additional spread, conservative advance rates, tight covenant triggers.
The math collapses below 75 obligors
NRSRO rating agencies fall back to individually-rated obligors when pool size is insufficient. That kills the economics for sub-IG operators. The 36-month / 100+ operator threshold is the floor.
Default? Severity assumed at 0%
Rating agencies need realized recovery data to model loss-given-default. Absent that, they assume zero recovery — devastating for asset-heavy operators where collateral recovery is the whole thesis.
Sixteen modules. One canonical graph. Hash-chained, time-stamped, attested at write.
Operators adopt Allometry's sixteen modules for urgent ops pain — margin leak, dispatch chaos, supplier concentration. Every commercial decision they make deposits a structured, hash-chained record into the operator's vault. The covenant → module → trigger map translates that record into the exact field set each lender tier enforces. ZKP attestation lets the operator prove the threshold is cleared without exposing the underlying data.
Six weeks of diligence collapses into seconds because the work is already done — continuously, by the model, before the lender ever opens the file. The lender's covenant slots arrive pre-filled. Verify the proofs, set the thresholds at close, decision.
See the covenant map → See the vault →The vault does not unlock tiers on a calendar.
It unlocks them as the covenant evidence each tier enforces reaches the vintage that makes that evidence statistically usable. Six months of attested ledger is enough for working-capital banks because their covenants are near-term and operational. Twelve months unlocks infrastructure debt because forward-looking SLA and throughput data become predictive. Twenty-four months unlocks insurance because failure-rate distributions stabilize. Thirty-six months unlocks pension paper because rating-agency loss curves finally have enough vintage to fit a granular-pool methodology.
The math is brutal and predictable, and it is also what Meta's TRIBE v2 result (May 2026) empirically validated: a foundation model trained on neural data from 700+ subjects produces a ~70× resolution lift over the same model class trained on 4 subjects. Pulse follows the same curve, applied to operators. Pulse at one design partner is the four-subject model. Pulse at 25 operators is T2-ready. Pulse at 100+ operators is TRIBE-grade, which is what unlocks rated paper. The graph is the dataset. The dataset is the moat. The vintage curve is the timeline of when each tier becomes addressable.
Four tiers. Four deep-dives.
Each tier page covers the named lenders with current 2025–2026 activity, typical facility size, diligence timeline, document requirements, key underwriting metrics, pricing reality, and exactly what the vault unlocks — quantified, with concrete spread/structure savings.
Banks · ABL · Working capital · Equipment finance
Live Oak, BDC, Siena, Pathward, Wells Fargo Capital Finance, RBCx, Crestmark. Collapse 6 weeks to days. Save 50–100 bps on pricing. Skip the field exam.
Infrastructure debt funds
Brookfield BID IV, Generate Capital, Stonepeak, BlackRock/GIP, Macquarie, Nuveen, Allianz. The 5th credit pillar — operator-as-guarantor.
Insurance-linked capital · parametric · performance bonds
Munich Re, Swiss Re, Hannover Re, Nephila, Markel, AXA XL, Parametrix, Descartes. Allometry as the actuarial table — and eventually the carrier.
Pension allocators · sovereign wealth · rated paper
CPPIB, CDPQ, CalPERS, APG, NBIM, GIC, ADIA. NRSRO-rated paper. The category endgame. KBRA-rated solar ABS is the precedent template.
Two front doors. One vault between them.
If you operate a $10–100M asset-heavy business and capital is gated by data fragmentation, run the Vault Readiness Audit. If you underwrite asset-heavy operators and want to pressure-test the covenant map against your portfolio, run a free Portfolio Blind-Spot Scan.