Loop Raises AI Platform To Predict Supply Chain Disruptions

Loop, a San Francisco startup led by founders Matt McKinney and Shaosu Liu, closed a $95 million Series C led by Valor Equity Partners and the Valor Atreides AI Fund, with participation from 8VC, Founders Fund, Index Ventures, J.P. Morgan Growth Equity Partners, and Tao Capital Partners. The company sells a verticalized AI platform that transforms fragmented operational data into predictive and prescriptive intelligence for logistics and supply chains.

Loop operationalizes enterprise data by combining internal and external models via a coordination layer the company describes as a harness. The platform ingests hard-to-parse inputs-scanned documents with no OCR text, PDFs, paper forms, and digital messages-and converts them into structured records to power workflows and forecasts. Loop cites its DUX family of AI models and agents as the core inference layer, while also orchestrating frontier models where appropriate. Key technical capabilities include:

• •automated extraction and normalization of unstructured logistics and finance data
• •a model orchestration layer that sequences specialized models and agents
• •connectors and data mappings for ERP, TMS, WMS, and order management systems

Supply chains are a difficult domain for AI because data is inconsistent, siloed, and operationally sensitive. Loop targets the high-value back-office surface where better visibility directly affects working capital and cost-to-serve. By focusing vertically, Loop reduces domain mismatch risk that generalist LLMs face and creates defensible integration work-data schemas, system connectors, and domain heuristics that are costly to replicate. The move from diagnostics toward predictive and prescriptive actions aligns with a broader trend: enterprises increasingly expect AI not only to surface anomalies, but to recommend concrete operational changes and to automate follow-up workflows.

If you operate data pipelines, MLOps, or supply chain systems, Loop exemplifies a production pattern where value requires more than a model. Implementation demands:

• •robust data ingestion and denoising for legacy formats
• •integration with transactional systems to close the loop on recommendations
• •model orchestration that mixes specialized and foundation models while preserving auditability and latency SLAs

Loop plans to invest the raise in product engineering and hiring, signaling accelerated work on connectors, model training for domain-specific tasks, and increased enterprise deployment support.

Watch for technical signals in Loop releases: detailed docs on DUX capabilities, published benchmarks for prediction accuracy and time-to-action, and new connectors for major ERP/TMS/WMS vendors. Also watch customer case studies that quantify savings and working capital improvements, and whether Loop exposes APIs or partner tooling that change integration effort for internal teams.

The funding validates that enterprise AI that solves integration and data-quality problems still attracts large venture capital. Loop is betting that verticalized model stacks plus deep system integration create defensible product-market fit in a domain where measurable financial impact is decisive.