The $2.5T number that won't move
The Asian Development Bank publishes the only global benchmark of unmet trade-finance demand. In their 9th iteration (released January 2026, data 2023-2025), they put the global trade finance gap at $2.5 trillion, unchanged from 2023, representing approximately 10% of global merchandise trade. The 2018 number was $1.5T. The 2020 number was $1.7T. The gap grew, then stuck.
The gap is the size it is because most of it can’t be cheaply underwritten. Trade finance supports 80% of global trade through letters of credit, trade loans, guarantees, and insurance, with bank-supported volume around $9 trillion annually. Trade finance is also one of the lowest-default asset classes in banking; ADB’s Trade Finance Register has documented this for ten years. The persistent gap is therefore not a credit-risk problem. It is a verification-cost problem.
Look at who gets rejected. The 2025 SME rejection rate is 41%, only just convergent with the corporate rejection rate of 40% (ADB notes the convergence is partial progress and requires more research). Women-led businesses see 70% of applications partially or fully rejected. Deep-tier suppliers and emerging-market corridors are concentrated where the rejection rates are highest. Pages 01 through 04 documented why this happens at the mechanism level. This page documents the size of what remains unfixed.
What digitization alone unlocks
Industry estimates of what becomes possible from going digital are substantial.
WTO Trade Facilitation Agreement implementation: $750 billion to $1 trillion per year in increased global trade from paperless customs and digital signatures. McKinsey on full eBL adoption: $40 billion in additional annual trade volume.DCSA: $6.5 billion in direct cost savings from the paper-to-eBL transition, plus $30-40B in trade growth. ICC: digitization could cut trade finance costs by $6 billion in three to five years and boost banks’ trade finance revenues by 10%.
These numbers are real. They are also bounded. All of them assume the trust primitive stays the same. Paperless trade with the same UCP 600 documents-alone mechanism is faster, cheaper, less prone to physical loss, more discoverable. It is not differently underwritable. The trade finance gap doesn’t close materially under digitization alone, because the verification cost per trade falls only modestly when the document changes from PDF to digital signature. The marginal SME rejected at 41% is still rejected after digitization, because the bank’s underlying constraint, cost-of-verification versus revenue-on-trade, hasn’t structurally changed.
What primitive replacement unlocks
Primitive replacement is what makes the structural shift.
Real-time verification flips the SME unit economics. Today’s KYC at $2,500 to $3,000 per corporate review (page 04) makes a $50K to $200K SME trade economically marginal to underwrite. With attestation-based verification, multi-party signatures generated as the normal output of operations, the verification cost per trade approaches zero, because the work is already done during the trade. The 41% SME rejection rate is a unit-cost problem that the trust primitive determines, not a credit problem.
Verified events unlock atomic settlement. Today’s LC cycle locks working capital for 30 to 180 days because the bank can’t act until documents-alone clear under strict compliance. With a verified milestone event signed by carrier, customs, and inspection, settlement releases automatically against verified completion: no T+2, no correspondent-banking reconciliation, no documents-alone discrepancy review. Hash-matched evidence decomposes corridor risk. An emerging-market trade today carries a country risk premium because the bank can’t verify what happened at the foreign port. With independent attestations from the carrier, customs, and inspection body, the corridor-specific risk decouples from the trade-specific risk. Banks price the trade, not the country.
- Verification cost$2,500–$3,000 per corporate KYC review
- Compliance ratio$4.04 spent per $1 of fraud loss prevented
- Cycle time30–180 days, documents-alone examination
- CounterpartyTop-tier corporates, established corridors
- Minimum economic trade$500K+ to make underwriting profitable
- Who’s rejected41% of SMEs, 70% of women-led firms, most deep-tier suppliers
- Verification costNear-zero per trade (attestations as byproduct of operations)
- Compliance ratioMulti-party signed evidence, no after-the-fact reconciliation
- Cycle timeReal-time settlement on verified milestone
- CounterpartySMEs, deep-tier suppliers, emerging-market corridors
- Minimum economic trade$10K+ becomes underwritable
- Risk decompositionTrade-specific risk priced independently from corridor risk
The RWA bridge
The capital to underwrite this newly addressable market is arriving faster than anyone expected.
Real-world asset tokenization tripled in 2025 alone, from $5.5B (January 2025) to $18.6B (December 2025), and reached roughly $30B by year-end. The market has grown 934% from $2.9B in 2022. Conservative forecasts (McKinsey, assets-only) put the 2030 number at $2 trillion. Broader forecasts (BCG + ADDX) put it at $16 trillion. Ripple + BCG project $18.9 trillion by 2033. Standard Chartered and Synpulse specifically project trade finance to reach 16% of the tokenized RWA market by 2034.
The regulatory infrastructure now exists. The GENIUS Act (July 2025) gave US payment stablecoins a federal framework. EU MiCA is already in force. MLETR-aligned legislation in the UK, Singapore, Bahrain, ADGM, India, and Netherlands gives electronic trade documents legal equivalence with paper. Institutional issuers are already shipping production-grade tokenized products: BlackRock BUIDL (over $500M in months), Franklin Templeton FDIT, Apollo ACRED, Fidelity Digital Interest Token (September 2025), Siemens €300M corporate bond on-chain, JPMorgan Kinexys settling Treasuries on a public chain with delivery-versus-payment.
Private credit is the proof-of-concept ($14-16B already tokenized, the largest non-stablecoin RWA segment). Trade-finance receivables are structurally the next wave: shorter duration than private credit, lower default rates per ADB, naturally fragmented (a $400K shipment is a tokenizable unit), and uncorrelated with public markets. The capital is looking for it.
- TRADE OS1Trade event happensSME runs the trade through TradeOS the way they’d run an ERP: quote, order, supplier, shipment, inspection, customs, delivery. The attestations are a byproduct of operating.Legacy: manual entry across 9–12 systems, 4% error rate, 11+ hours/week per team.
- EDMA L22Multi-party attestationsCarrier signs cargo loaded. Inspection body signs goods verified. Customs signs cleared. Each signature is a hash of the underlying evidence, scoped to a specific event.Legacy: paper bill of lading, paper inspection cert, paper customs filing, trust by paper as documented in page 04.
- EDMA L23Proof of Verification consensusAttestations hash-match against the same evidence. Network commits exclusively once. Duplicate financing becomes structurally impossible, not retrospectively detectable.Legacy: Hin Leong borrowed against the same cargo six times because no one could see the others.
- MARKETPLACE4Verified receivable listedThe verified trade-finance receivable appears in the marketplace. Financiers see the cargo, the counterparty history, the milestone schedule, and the verifier identities.Legacy: the trade would have been rejected before reaching this stage (41% SME rejection).
- MARKETPLACE5Financier funds in EDSDFinancier commits stablecoin liquidity. The commitment is escrowed against verified milestone delivery. EDMA captures a 0.25% fee on financed principal at this step.Legacy: SBLC at 0.5–4% of LC value, 6–12% APR on lines of credit, 30–180 day cycle.
- SETTLEMENT6EMT milestone releaseAs each verified milestone fires (cargo loaded, customs cleared, delivery confirmed), the corresponding EMT releases the committed funds in tranches. The smart contract enforces the schedule.Legacy: bank holds funds until documents-alone clears, with strict-compliance discrepancy reviews.
- SETTLEMENT7Atomic settlementFunds move on EDMA L2 in seconds. No correspondent-banking chain to reconcile. The financier earns yield on a verified, exited receivable. The SME has working capital free for the next trade.Legacy: T+2 at minimum, multi-bank SWIFT reconciliation, manual back-office workflow.
What EDMA captures
EDMA is the four-layer stack that delivers verified trade events to underwriting capital.
TradeOS is the operations layer. SMEs run their trade through it the way they would run an ERP: quote, order, supplier, shipment, customs, inspection, payment. As a byproduct of operating, TradeOS generates the attestations the protocol needs. Proof of Verification is the consensus layer on the EDMA L2: multi-party attestations are hash-matched against the same evidence and exclusively committed once network-wide. Settlement is the value layer: EDSD stablecoin for commitment, EMT milestone tokens for release-against-verified-completion. The Marketplace is the capital layer: financiers underwrite verified trade-finance receivables in EDSD, earn yield, and exit on settlement.
The capture model is not winning a slice of the existing $9 trillion bank-supported trade finance market. It is expanding the addressable market into the $2.5 trillion unmet segment: the SME trades banks today can’t profitably underwrite, the emerging-market corridors where country risk overprices trade-specific risk, the deep-tier suppliers who never reach the bank’s underwriting screen.
EDMA Group ran physical trade for six years on the broken stack. We were the kind of SME the 41% rejection rate is about. We watched LCs cost 4 to 7% all-in, watched compliance teams examine documents for two weeks, watched a $400K shipment turn into 90 to 180 days of locked working capital. We watched our bank ask us to translate the same shipment data nine ways across nine systems. We built TradeOS because we needed it. We built the marketplace because the financiers we needed didn’t exist for trades our size. We built the L2 because the trust primitive on top of which all of this depends is the piece that has to change for the rest of the stack to make economic sense. The thesis isn’t theoretical. We are also the first customer.
Why now
The legal, financial, operational, and regulatory preconditions converged in 2024-2025. They did not exist as a coherent stack in 2022.
Legal: MLETR-aligned legislation in the UK (2023), Singapore, Bahrain, ADGM, India (2025), Netherlands (2025); pending in Germany, France, Japan; AfCFTA Digital Trade Protocol committed Africa-wide. Financial: RWA tokenization at $30B and tripling year-over-year, GENIUS Act and MiCA establishing stablecoin legal frameworks in major jurisdictions, institutional issuers shipping production-grade products. Operational: DCSA’s 9 ocean carriers (70% of global container trade) committed to 50% eBL adoption by 2027 and 100% by 2030, first cross-platform eBL interoperability achieved May 2025. Macroeconomic: 80% of polled banks (ADB 2025) expect rising trade finance demand from supply chain realignment and corridor diversification.
We have spent two years building for this moment. Trade OS Platform is in production and onboarding 50 beta customers through July 2026 (Stage 1). L2 testnet hardens through August 2026; mainnet promotion in October 2026, with the first $EDM protocol burns going on-chain (Stage 2). Global Trade Marketplace launches November 2026 through January 2027: commodities RFQ marketplace with EMT milestone-gated settlement, replacing letters of credit on five lanes from supplier-finder to settlement (Stage 3). $EDM DEX listing on Uniswap v3 follows in Q1 2027 at a $1.00 target price with $18M locked liquidity. The presale is open now. Full roadmap at /roadmap/.
