Logistics · Deep dive

Loop

AI-native freight audit and payments — a platform that ingests every invoice, bill of lading and contract in any format, audits 99% of freight bills without human touch, pays carriers through J.P. Morgan rails, and is now repositioning as the 'intelligence layer' for enterprise supply chain spend.

emerging

The question that decides it: Loop books revenue as a tiered percentage of the freight payment volume it audits, and Valor priced the April 2026 Series C on the promise that DUX's contract-and-invoice data compounds into the supply chain's 'intelligence layer.' The mechanism in question: does audit-grade structured data — contracts, invoices, BOLs normalized across ERP, TMS and WMS — actually generate predictive spend intelligence customers pay for beyond the one-time 2-7% audit savings, before frontier-model document extraction commoditizes the ingest layer and lets Cass, U.S. Bank and the TMS vendors give away 99% no-touch audit at bank-subsidized, pennies-per-invoice pricing a venture-backed take rate cannot undercut?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Loop applies AI to the least glamorous corner of US logistics: checking freight bills and paying them. Founded in 2021 by two architects of Uber Freight, it ingests invoices, bills of lading and contracts in any format, audits them against digitized contract terms — 99% of invoices with no human touch, the company claims — and executes payment, with a June 2024 J.P. Morgan Payments partnership adding same-day carrier quick-pay. It has raised roughly $160M, most recently a $95M Series C led by Valor Equity Partners in April 2026, with customers from GILLIG and Great Dane to Loadsmart. It matters now because the Series C reprices it from freight-audit tool to would-be “intelligence layer of the entire supply chain” — a claim to be tested against century-old bank incumbents on one side and commoditizing document AI on the other.

Founding story

Matt McKinney and Shaosu Liu met at Uber and were both architects of Uber Freight — McKinney a data science manager on the brokerage’s pricing and launch, Liu the engineering lead who took the product from zero past $2B in annual revenue. From inside a digital brokerage they watched the industry’s financial plumbing fail at scale: roughly 20% of freight invoices carry errors, an invoice takes about 50 days on average to pay, and shippers leak money on charges that never matched the contract. The insight: this is a data problem — the source of truth for what a shipper owes is scattered across emailed PDFs and contract appendices nobody reads — and legacy audit firms, paid per invoice or on gain-share, had little incentive to fix it.

They built the prototype on nights and weekends, left Uber in May 2021, and raised a $6M seed co-led by Susa Ventures and 8VC. The cap table is a who’s-who of their old world: Uber co-founder Garrett Camp invested through Expa, early Uber employee Ryan Graves through Saltwater Capital, and Flexport’s Ryan Petersen came in early; the engineering team drew from Uber and Flexport. Loop came out of stealth in November 2022 with $30M across seed and a Founders Fund-led Series A. Founded and long headquartered in Chicago, it now presents as San Francisco-based — a quiet migration toward the AI talent pool the Series C is earmarked to hire from.

How it works

The core machine is DUX, Loop’s family of document-and-domain models. Carriers send invoices in whatever form they like — EDI feeds, emailed PDFs with no machine-readable text, paper scans — and DUX ingests, normalizes and links each to the underlying shipment record and governing contract before any audit logic runs. That linking step is the differentiator: a fuel surcharge can only be verified against the fuel table in that customer’s contract for that lane, so Loop digitizes the contracts themselves, then audits every invoice line — rate, accessorials, service failures, duplicates — against customer-specific controls. Exceptions route to a human-in-the-loop queue with root cause attached; disputes go back to carriers with documentation.

Downstream, the same workflow handles GL coding, cost allocation and payment execution — audit and pay as one pipeline, not two vendors. The J.P. Morgan Payments integration (June 2024) turns that pipeline into a working-capital product: approval compresses from roughly a week to about four hours, and approved carriers opt into quick-pay, taking same-day payment on J.P. Morgan rails for a pre-negotiated discount — the shipper pockets the discount, the carrier gets cash flow, and Great Dane runs 15% of its full-truckload payments through the program. Under the hood, Liu describes a harness coordinating in-house logistics-document models with frontier models, and the Series C-era platform is extending ingestion into supplier, customs, warehouse and procurement data via ERP, TMS and WMS integrations.

Product and business overview

The platform breaks into sellable layers. Freight and parcel audit covers every mode — truckload, LTL, parcel, ocean, air — with contract digitization and no-touch audit. Payment automation adds carrier invoice pay, GL coding, cost allocation and the J.P. Morgan quick-pay program. Spend intelligence is the analytics tier: cost-to-serve visibility, root-cause analysis of overcharges, and benchmarking ammunition for contract renegotiation — marketed to CFOs as “Freight Audit 2.0” with claimed 2-7% savings. In May 2026, three weeks after the Series C, Loop packaged the stack as the Logistics Data Platform, powered by DUX 2.0 — an AI-native supply chain data system rather than a FAP vendor. It sells to both sides of the market: shippers (GILLIG, Great Dane, Dot Foods, Olipop, Kendra Scott, Clemens Food Group, Outset Medical) and logistics providers (Loadsmart; Convoy until its October 2023 collapse). A Princeton TMX partnership embeds audit-and-pay inside that TMS, and JPMorgan Chase itself is a named customer — the partner-as-customer-as-investor triangle that defines the company.

Business model and pricing

Loop publishes no rate card. Third-party analysis (Thirdfin, 2026) and Loop’s own marketing describe a tiered, consumption-based model: revenue is a fixed percentage of total payment volume through the platform, declining as volume grows. That deliberately breaks with legacy FAP economics — per-invoice fees plus gain-share, where auditors take 5-50% of savings found and are paid more when contracts stay messy. The consumption model aligns Loop with volume, not errors, and makes revenue a function of freight rates and tonnage — meaning the 2022-2025 freight recession was a direct headwind to its own top line, a sensitivity no press release mentions. Quick-pay adds a second engine: the spread on early-payment discounts, shared among shipper, Loop and J.P. Morgan. No ARR has ever been disclosed; the only public volume marker remains the $3B in booked total payment volume cited in October 2023.

Traction over time

The gap in the series is conspicuous: no payment-volume, customer-count or revenue figure has been published since October 2023, through two subsequent fundraises. Headcount of ~160-220 against ~$160M raised implies a company still investing well ahead of revenue.

Market analysis

The honest market is small and the aspirational one enormous. Mordor Intelligence pegs the global freight audit and payment market at $0.97B in 2025, growing at a 14.2% CAGR to $1.89B by 2030; other 2024 estimates range from $285M to $3.2B — a reminder that FAP as a service category is a rounding error against the spend it polices. The pool Loop actually monetizes is the payment volume itself: transportation and logistics exceed 8% of global GDP (US BTS, cited by Loop in 2023), and a take rate on audited freight spend scales with that, not the FAP services market. Structural forces favor the pitch: 2025-26 tariff volatility forces enterprises to understand landed cost per shipment, 20% invoice error rates persist, and recession-squeezed CFOs will buy anything credibly promising 2-7% back. The countervailing force is the one lifting Loop: frontier models are making document extraction — the hard thing DUX does — cheaper for everyone, incumbents and in-house teams included.

Competitive intel

Loop fights on two fronts. Above it sit the bank-model incumbents: Cass Information Systems, a NASDAQ-listed bank holding company founded in 1906 that claims over $50B in freight processed annually, and U.S. Bank Freight Payment — both earning float on payment balances, letting them price audit at levels a venture-backed take rate struggles to meet, and both owning the Fortune 500 treasury relationships. Beside it sit the FAP specialists: Trax Technologies (enterprise, global, now AI-branded), Intelligent Audit (25+ years, per-invoice pricing), CTSI-Global and nVision Global (service bureaus with decades of carrier EDI plumbing; CTSI bundles a TMS, which Loop does not). Below it, AI-native insurgents like Freehand and AP-automation vendors such as OpenEnvoy make Loop’s own argument against it — Freehand specifically attacks Loop’s human-approval exception queues as half-automation. Loop’s edge is genuine: a 2026 Gartner Market Guide listing, contract-digitization-first audit that catches what legacy providers pass through, and a J.P. Morgan triangle no rival can copy. Its exposure: every layer of its stack — extraction, audit rules, payment rails — has at least one competitor doing that layer at larger scale or lower cost.

History and evolution

• May 2021 — Founded in Chicago at the peak of pandemic supply-chain chaos; $6M seed co-led by Susa and 8VC.
• Nov 2022 — Exits stealth with $30M (Series A led by Founders Fund); positioning: “logistics payments platform.”
• Oct 2023 — $35M Series B co-led by J.P. Morgan Growth Equity Partners and Index; $3B+ YTD booked payment volume. The same month, flagship customer Convoy collapses — a logo gone within weeks of the press release naming it.
• Jun 2024 — J.P. Morgan Payments partnership; working-capital and quick-pay products launch; positioning: “modern audit & payment platform.”
• 2025 — Freight recession grinds on; CFO-facing “Freight Audit 2.0” messaging and TMS partnerships (Princeton TMX); no metrics disclosed.
• Apr 2026 — $95M Series C led by Valor Equity Partners and its Atreides AI Fund; positioning: “full-stack, verticalized AI platform.”
• May 2026 — Logistics Data Platform and DUX 2.0 launch; ambition restated as the supply chain’s intelligence layer.

Three brand identities in five years — payments platform, audit-and-pay platform, verticalized AI platform — track capital-market tastes precisely: either savvy narrative management or a company still searching for the durable version of itself.

What people say

The case for. The named-customer evidence is specific and quantified. GILLIG’s VP of aftermarket parts credited Loop with 6% transportation savings and a move from gut-checking 30% of invoices to full automation (Oct 2023). Great Dane reported meaningful margin recapture on quick-pay (Jun 2024). A Fortune 500 customer found 2% of spend in year one that its legacy FAP provider had missed. Valor’s Antonio Gracias — whose fund did what Liu called “very deep diligence” on defensibility — argued Loop turned the hardest data in the supply chain into durable advantage; Thirdfin’s 2026 review calls it one of the most technically ambitious platforms in the category.

The complaints. The independent review surface is close to empty, which is itself the finding: Thirdfin (Jun 2026) confirms no G2 or Capterra scores and zero Gartner Peer Insights reviews despite the Market Guide inclusion — five years in, essentially every public proof point routes through the company or its investors. (Glassdoor is unusable: the several “Loop” employers on the platform, including one carrying scam allegations, are different companies — a name-collision problem that also muddies diligence against Loop Returns, the ecommerce firm.) Loop’s own marketing statistics wobble — 20% invoice error rates in some tellings, 80% in the 2022 TechCrunch coverage — inconsistency that invites CFO skepticism. And the traction silence since October 2023, through two raises and a freight recession that mechanically shrinks consumption-based revenue, is the complaint the market registers by omission.

Outlook: the open question

For Loop to work, audit exhaust has to become an intelligence product customers pay for on its own — concretely: the ERP/TMS/WMS-connected data layer launched in May 2026 must drive renewals and expansion after the one-time 2-7% audit savings are harvested, and the take rate on payment volume must survive bank incumbents whose float income lets them price audit near zero. The bull case is real: the founders built Uber Freight’s actual systems, contract digitization demonstrably catches money legacy audit misses, the J.P. Morgan triangle is distribution no startup rival can copy, and Valor’s diligence-heavy check suggests someone with frontier-lab access judged the moat defensible. If DUX’s audit-grade dataset genuinely compounds, Loop becomes the system of record for logistics spend the way Cass never bothered to.

What would sink it: the ingest layer commoditizes faster than the intelligence layer materializes. McKinney admits the founding assumption was that the enabling AI wouldn’t arrive until ~2030; it arrived early, for everyone. If frontier models make invoice extraction a solved problem, Cass and U.S. Bank bolt adequate AI onto float-subsidized pricing, TMS vendors bundle audit as a checkbox, and Loop is left selling dashboards against a shrinking wedge — three rebrandings in five years suggesting the company already feels that clock. The tells: whether Loop ever publishes a payment-volume or ARR figure again (silence since Oct 2023 is the loudest number it has); whether Logistics Data Platform deals close with buyers outside the freight-audit budget line; whether quick-pay grows into a real financing business; and whether the next round is priced up by an outsider or defended by insiders. Loop has the right founders, partner and wedge — what it has not yet shown, anywhere in public, is that the wedge converts.

How a challenger would attack it

Loop is the challenger, so the attack comes from a leaner copy running Loop’s own thesis to its conclusion. McKinney concedes the enabling AI arrived years early — for everyone — which means DUX’s hard-won extraction layer is exactly what a 2026 entrant gets nearly free from frontier models. Freehand already shows the angle: market autonomous exception resolution against Loop’s human-approval queues, framing Loop as half-automation with a services tail. The pricing attack writes itself. Loop’s take rate is a percentage of total payment volume — revenue that shrank through the freight recession and scales with the very spend it promises to cut — so a challenger sells flat per-invoice or seat pricing and lets CFOs keep 100% of the 2-7% savings, undercutting the take-rate math the same way the file says bank float already does. The proof-point vacuum is the softest tissue: five years in, Loop has zero G2, Capterra or Gartner Peer Insights reviews, no volume or ARR disclosure since October 2023, and marketing error-rate claims that wobble between 20% and 80%; a challenger that publishes audited case studies and transparent pricing wins every skeptical CFO evaluation. And three repositionings in five years hand the attacker its narrative: pick one buyer — the freight-audit budget line — and be boringly excellent there while Loop chases the “intelligence layer.”

Same playbook, new buyer

The playbook — digitize contracts, link every invoice line to the governing agreement, audit no-touch, pay on bank rails — is not freight-specific; freight is just where McKinney and Liu saw it first. The nearest adjacency is other high-error, contract-governed B2B spend: ocean demurrage and detention, customs brokerage fees, warehouse and 3PL billing — Loop gestures at these with DUX 2.0 but its models, carrier integrations and J.P. Morgan program are all built around US domestic freight. Further out, the same architecture fits utility, telecom and healthcare-adjacent spend categories where invoices routinely diverge from rate schedules and incumbent auditors work on gain-share, the misaligned model Loop’s consumption pricing was designed to kill. The other axis is buyer size: Loop sells enterprise deals to GILLIG-class shippers, leaving the mid-market shipper with $5-50M of freight spend — too small for Cass, too error-prone to ignore — unserved; a self-serve product at that tier builds volume the bank incumbents’ service-heavy model can’t chase. Loop won’t follow down-market soon: ~$160M raised against 160-220 heads demands enterprise ACVs, and its Valor-priced story points up the stack, not down.

Sources and further reading

• Loop lassos ex-Uber talent and money to finally fix freight invoicing (TechCrunch, Nov 2022)
• Loop breaks into logistics payments space with $30M in early funding (FreightWaves, Nov 2022)
• Breakthrough Transportation Logistics Startup Loop Announces $35 Million Series B (BusinessWire, Oct 2023)
• Loop and J.P. Morgan Payments to Improve Access to Working Capital Across the Supply Chain (BusinessWire, Jun 2024)
• Exclusive: Loop raises $95M to build supply chain AI that predicts disruptions (TechCrunch, Apr 2026)
• Loop Raises $95M Series C to Scale Its AI Platform Across the Supply Chain (BusinessWire, Apr 2026)
• Loop Freight Audit Review — Features, Pricing, Alternatives (Thirdfin, 2026)
• This Tech Could Make Complex Freight Billing a Thing of the Past (Sourcing Journal, Oct 2023)
• Freight Audit and Payment Market Size & Growth to 2030 (Mordor Intelligence, 2025)
• Freight Audit 2.0: The CFO’s Guide to Unlocking 7% Savings in Transportation Spend (Supply Chain 24/7, 2025)