One sovereign L1. One kernel. Eight industries.
JIL Sovereign is not eight platforms. It is one settlement engine, configured eight ways. The kernel predates these verticals by 18+ months. Each new industry is a check pack on top of the same engine - not a new platform. That asymmetry is the moat.
We didn't build eight platforms. We built one and configured it eight ways.
Most "blockchain for compliance" startups fork Ethereum, repackage it, and try to bolt regulation on top. We didn't. We built a sovereign L1 from line one - our own consensus, our own ledger, our own validator admission, post-quantum signing in production today. ~1.4 million lines of original code. The kernel is 18+ months old.
That kernel produces a Court Ready Evidence Bundle (CREB) for every interaction. Same output format whether the input is a Medicare claim, a tokenized-Treasury settlement, an EB-5 source-of-funds document, an H-1B Labor Condition Application, or a $10K USDC transfer. The verticals are testimony to the modularity, not separate products.
Build-the-kernel vs add-a-vertical.
The economics of the platform come from the asymmetry. The kernel is built once. Verticals ride on top. The cost to add the next vertical is a fraction of the cost to add the first one - because the first one paid for the kernel.
Sovereign L1 + attestation network
- 14-of-20 BFT consensus across 13 jurisdictions
- Post-quantum signing (Ed25519 + Dilithium-III hybrid) in production
- HSM-backed validator keys, 7-day rotation
- Sealed CREB output, FRE 902(14) self-authenticating
- 148-check attestation catalog, MPC custody, BIP-44 multi-chain
- Kafka + Redpanda settlement consumer, Postgres ledger router
- Orchestrator + tier1-pipeline + ava agentic-AI trio
- 53 patent claims (20-30 net-new in active drafting)
Customer-profile-LOB-gated check pack
- ~3 check definitions (SQL aggregates over a vertical schema)
- 1 public-data ingest script (federal CSV / API / blockchain RPC)
- 1 template pack (3 narrative + 3 remediation, gated on lob)
- 1 schema migration (vertical-specific tables + finding store)
- 1 proof page on the public site, sample CREB
- ~12-15 files total, 1 PR, deployable inside a session
A competitor entering any of these eight industries must first build a kernel that took us 18+ months. We add new industries in days. That asymmetry is the platform value, and it compounds with every additional vertical we ship.
What's underneath every CREB we seal.
The kernel is the part the verticals do not touch and never replace. It is the reason a Medicare-claim CREB and a tokenized-Treasury CREB are evidentiarily indistinguishable to a court, regulator, or auditor.
Sovereign Compliance Network (SCN)
10 mainnet validators across 13 jurisdictions today. 14-of-20 BFT quorum. Every validator is a named legal entity with a documented jurisdiction, HSM-backed key custody, and post-quantum signatures. Anonymous validators are not permitted by design.
Post-quantum from line one
Ed25519 + Dilithium-III hybrid signing in production today, not on a roadmap. Kyber KEM in production. ML-DSA-65 implemented. Every key is generated and held inside an HSM and rotated on a 7-day cadence. A record sealed on JIL in 2026 will still be verifiable in 2050.
CREB sealed receipt
Every check, every verdict, every finding ships as a Court Ready Evidence Bundle. FRE 902(14) self-authentication, 15+ year retention, content-addressable hash, full reproducibility manifest (code version, query plan, signal thresholds). Independently verifiable without JIL's cooperation.
Ava: in-house agentic AI
Tier 1 verdict path is deterministic - SQL aggregates only, no stochastic LLM. Ava groups, narrates, and routes findings; never produces the underlying flag. Numeric-drift-guarded LLM refinement; 4-way AND-gate prevents accidental GPU spend on POC engagements. No external LLM API - JIL operates the model directly.
Every vertical is a configuration, not a product.
Each card below points to the live public-data POC for that industry. Same kernel underneath all of them. Same CREB output. Same Ava agentic layer. Different check packs, different schemas, different ingest sources, different LOB gating.
Capital Markets Settlement Integrity
14,156 chronic-offender CUSIPs surfaced from 339,992 SEC fails-to-deliver records. Top: SPY $13.0B, IWM $8.76B, LQD $7.46B.
Federal Grants Integrity
1,000 CY2024 federal awards / 321 recipients / 15 agencies. 66 concentration findings (39 critical). California DHCS 21.36% of HHS, Amtrak 39.33% of DOT.
H-1B Wage & Sponsor Integrity
FY2024 Q3+Q4 LCA disclosures, 47,132 distinct sponsors. 459 findings: 134 wage-below-prevailing, 250 employer-concentration, 75 NAICS anomaly.
P2P Settlement Integrity
1,000 USDC transfers from Etherscan + 46 sanctioned-address seed (Tornado Cash, Lazarus, Hydra, Garantex). 38 findings.
EB-5 Investor Visa Integrity
USCIS Data Hub processing-times + regional center list. 25 findings: 2 terminated regional centers with recent processing activity, I-526 denial-rate spike, 22 long-pending I-829 cohorts.
Trade Finance Integrity
924 trade flows / 527 corridors. Cross-referenced against 8 OFAC programs (IRN, PRK, CUB, SYR, RUS, BLR, VEN, MMR). 450 findings.
P&C Insurance Integrity
FARS 2022 National CSV. 450 findings: VIN-repeat clusters, geographic crash clustering at 500m, make/model fatal-rate outliers.
Workers' Compensation Integrity
1,848 NAICS-state-year cohorts indexed. 207 findings: 124 NAICS outliers, 83 state-concentration patterns. CA holds 65% of NAICS 62 health-care DAFW.
~1,945 findings across 740,000+ real federal and on-chain records. All produced on day one of each engine. All sealed as CREB records. All replayable bit-identically by a third party with the same source files and the manifest. The economics are not hypothetical.
The 18 months of work that the verticals ride on.
The verticals are recent. The kernel is not. The work below is what makes a one-week vertical add-on possible.
Pick the vertical you need today. Add the next one when you need it.
The customer engagement model follows the leverage: you commit to the vertical you actually need this quarter (one of the eight live), and adding a sibling vertical later is cheaper than your initial deployment because the kernel is already in place. Same applies to your industry's adjacent line of business - bringing it onto JIL is a check-pack add, not a re-platform.
Pricing reflects this. The four-SKU model (Tier 1 public-data baseline / Tier 2 commercial enrichment / Tier 3 court-ready CREB / Sovereign Stack in-tenant) prices the kernel once and the verticals on top.