Reducto Raises $75 Million Series B Led by Andreessen Horowitz to Power AI Document Intelligence Infrastructure
Reducto, a San Francisco-based startup at the intersection of AI and document intelligence, has raised US $75 million in a Series B funding round, bringing total funding to approximately US $108 million. The round was led by Andreessen Horowitz (a16z), with participation from existing investors including Benchmark, First Round Capital, BoxGroup, and Y Combinator.
Founded in 2023 by co-founders Adit Abraham (CEO) and Raunak Chowdhuri (CTO), Reducto offers an AI-powered platform that converts complex unstructured documents—such as PDFs with tables, redlines, charts and multi-entity layouts—into structured, LLM-ready inputs. The company combines traditional Optical Character Recognition (OCR) technology with vision-language models (VLMs) to enable machines to “read” documents in a more human-like way.
The Series B comes after a US $24.5 million Series A round earlier in 2025, which itself followed a seed round of US $8.4 million in late 2024. The latest financing will be used to accelerate model research, enhance product capabilities, scale infrastructure and increase adoption across both enterprise clients and next-generation AI teams. According to the company, monthly processing volume has grown more than six-fold since the Series A, now nearing one billion pages per month for leading AI-native startups, global financial institutions and a Fortune-10 enterprise.
Reducto positions itself as the “ingestion layer” for AI workflows: while much attention in the AI community focuses on large language models and generative applications, Reducto argues the real bottleneck is getting the raw document data into the right format. Its roadmap includes capabilities for document splitting (breaking large composite files into semantically meaningful chunks), extraction (mapping into structured schemas) and editing (allowing auto-fill or correction of documents via API). The company also announced flexible pricing options for early-stage teams and researchers, making its infrastructure accessible beyond just large enterprises.
Investor enthusiasm for Reducto underscores a broader trend in enterprise AI: the rise of infrastructure and data-preparation platforms that address foundational but often overlooked workflows. The lead investor, Andreessen Horowitz, described Reducto as poised to become “the indispensable infrastructure layer for AI” in document-intensive industries. Benchmark, a long-time supporter, said the company is “at the inflection point” where technology, demand and opportunity converge.
Despite strong momentum, Reducto faces the inherent challenges of the enterprise-software landscape: the global market for document-processing and enterprise ingestion infrastructure is populated with legacy providers and large cloud-platform incumbents, meaning Reducto must continue to prove its scalability, reliability, and ability to win long-sales-cycle major accounts. Equally, while parsing and ingestion are necessary for AI pipelines, the value-capture and monetisation path can be more complex than purely front-end applications.
Reducto’s co-founders bring impressive technical backgrounds (both MIT-educated and formerly at Google and NVIDIA) and the company emphasizes its early traction with customers in the AI, financial, healthcare and legal sectors, where massive document workloads remain a strategic bottleneck. Its claim to have processed hundreds of millions of pages for thousands of companies signals strong early validation. The company’s name and mission reflect the idea of breaking down complexity—cutting dense document stacks into clean structured data that high-performance AI can act on.
With the US $75 million Series B now secured, Reducto is focused on scaling its engineering, go-to-market and international expansion efforts, deepening its model research and expanding the platform into adjacent workflows beyond ingestion. The raise marks a significant milestone not just for Reducto, but for the enterprise-AI infrastructure category as a whole, signalling elevated investor confidence in companies building the “plumbing” of AI rather than just the visible applications.
