Loop Payments Inc. secured $95M in Series C funding to expand DUX, its AI model family that audits supply chain invoices by reading document structure, not just text.
Loop Payments Inc. closed a $95 million Series C on Thursday, pulling in Valor Equity Partners, the Valor Atreides AI Fund, and J.P. Morgan Growth Equity Partners to scale a platform that audits freight invoices using artificial intelligence. The company says its system compresses a process that typically consumes several weeks of manual labor down to roughly two hours.
The problem Loop is selling against is more common than it sounds. When manufacturers order parts or retailers book container space, the resulting invoices frequently carry billing errors that pass unnoticed until an audit team catches them, sometimes weeks after the fact. Those mistakes translate directly into unnecessary costs, and the document volume in logistics makes manual review both expensive and slow. According to SiliconAngle, Loop has built a family of AI models called DUX to automate most of that checking.
The freight audit problem
DUX is built on what Loop describes as a custom architecture tuned specifically for physical supply chain documents. The key design choice is that the models extract not just text but also structural signals: the positioning of form fields, stamps, and layout elements that carry meaning in logistics paperwork but that general-purpose document models tend to ignore. Loop says this spatial awareness lets DUX interpret invoices with greater accuracy than standard OCR or language model approaches.
After extracting data from a batch of invoices, DUX normalizes the information into a consistent schema and deploys AI agents to flag cost discrepancies. The platform also processes bills of lading, which confirm that a carrier has taken custody of cargo, and rate tables, the multi-variable pricing schedules that underpin most freight contracts and are notoriously difficult to audit manually. As SiliconAngle reported, Loop claims the full audit cycle shrinks from several weeks to about two hours.
Parcel tracking is a secondary capability. As DUX reads shipping documents it extracts location waypoints and feeds them into a visibility layer, giving supply chain teams a way to spot bottlenecks before they cascade into delays. The company has not published third-party benchmarks for either the speed or accuracy claims.
What the investors signal
J.P. Morgan Growth Equity Partners joining the round is worth noting. The bank's trade finance and logistics divisions handle substantial freight invoice volume themselves, which gives its investment a strategic angle beyond a pure financial return. Loop has not announced a commercial relationship with the firm, but its presence on the cap table adds weight to the company's enterprise sales story.
The broader backdrop is a wave of capital flowing into vertical artificial intelligence applications, where startups train narrow models against specific document types rather than building general-purpose assistants. Logistics is a natural target: processes are paper-heavy, error rates are quantifiable, and cost savings from catching billing mistakes are measured in dollars. That legibility makes the return-on-investment conversation with buyers far shorter than in use cases where AI benefits are harder to count.
Loop is not operating without competition. Established freight audit and payment processors have been layering machine learning onto their workflows for years, and larger supply chain software vendors have announced AI-driven invoice capabilities. What Loop is betting on is that document-architecture specificity, training models to read layout signals rather than content alone, is a durable technical advantage. That bet is plausible but unproven; the artificial intelligence review literature has not yet produced authoritative benchmarks comparing specialized logistics models against general document AI at scale.
At $95 million, Loop has runway to expand model training, broaden the document types DUX supports, and endure enterprise sales cycles that can stretch a year or more. Whether that is enough to hold a defensible position as foundation models improve and larger platforms absorb document-AI into their supply chain suites is the question its investors are paying ninety-five million dollars to answer.
Frequently Asked Questions
What does Loop's DUX model actually do?
DUX is a family of artificial intelligence models trained on supply chain documents. It extracts both text and structural layout information, normalizes the result into a standard schema, and uses AI agents to flag pricing discrepancies in freight invoices, bills of lading, and rate tables.
Who led Loop's Series C?
Valor Equity Partners and the Valor Atreides AI Fund co-led the $95 million round. SiliconAngle reported that J.P. Morgan Growth Equity Partners was among the institutional co-investors.
How fast does Loop claim to audit invoices?
The company says its platform compresses freight expense audits from the several weeks typical of manual review down to approximately two hours. No independent benchmarks have been published to verify that figure.
What is Loop's biggest competitive risk?
Established freight audit firms and enterprise software vendors are building similar AI capabilities. Loop's advantage depends on whether training models specifically for document layout signals remains meaningful as general-purpose foundation models continue to close the gap on specialized tasks.