Valon Raises $150 Million at $2.3 Billion to Put AI Agents Inside America's Mortgage Back Offices

Valon Technologies, a New York-based company that brands itself the AI-native operating system for regulated finance, announced Monday that it has raised $150 million in Series D funding at a $2.3 billion valuation — double its previous mark, per the company's release. The round brings in new investor Ribbit Capital alongside continued participation from existing backer Andreessen Horowitz.

The capital is aimed at one specific migration: moving the largest US mortgage servicers off mainframe-era systems and onto ValonOS, the company's platform for loan data, investor reporting, operational workflows, compliance logic and money movement. On top of that single system of record, Valon deploys AI agents that answer homeowner emails, allocate payments and run escrow analyses — with what the company describes as deterministic actions and an audit trail behind every step. Hiring is planned across engineering, product, deployment and go-to-market in New York and San Francisco.

The traction figures are company-reported and not independently audited, but they are unusually specific. Valon says it signed more than $200 million in contracted annual recurring revenue within six months of opening ValonOS to the industry, and that servicers responsible for one in six outstanding US mortgages are under contract to run on the platform. Two of the ten largest US servicers are already live: ServiceMac, the fourth-largest residential subservicer, and Carrington Mortgage Services, which acquired Valon's own servicing business in August and adopted ValonOS as its core platform. Rithm Capital's Newrez is also a client, though National Mortgage Professional notes that its migration — covering more than 4 million homeowners — is announced for 2027, so contracted volume is not yet live volume.

That distinction between "under contract" and "running in production" is the honest caveat in an otherwise strong story, and Valon's own history explains why the gap exists at all. Rather than selling software into an industry it had never operated in, the company spent six years running a fully licensed national servicer on its own platform — Valon Mortgage — to prove the technology could handle government and private-label servicing at scale. The Carrington acquisition in August let it exit first-party servicing entirely and become a pure software company, with the caveat that its most demanding customer is now also its former self.

The pitch to investors is that regulated finance is where AI agents get genuinely hard — and therefore where the winners get durable. "The bottleneck for deploying AI agents into regulated industries is context, not intelligence," co-founder and president Linda Du said in the release, pointing to the servicing data ontology Valon built while operating its own book. Andreessen Horowitz general partner Angela Strange framed the market as "$13 trillion" and argued Valon is positioned to repeat the playbook in other asset classes.

Valon says it plans to expand the same architecture into commercial, personal, auto and student lending, where high-volume transaction processing and strict regulation create similar constraints. Mortgage servicing has been the industry's most stubborn software backwater for sixty years; if the AI-agent wedge works here, the regulated back office — not the consumer chatbot — may turn out to be the quieter, stickier market for enterprise agents.