Mortgage servicing — the business of managing homeowners’ monthly loan payments — has run on aging mainframe systems for nearly 60 years in the US. Every regulatory change has piled on more technical debt and cost. “For sixty years, mortgage servicing has run on aging mainframe systems, and every regulatory change has compounded technical debt and increased costs,” said Andrew Wang, co-founder and CEO of Valon. “That is no longer the only option.”

Valon said it has raised $150 million in a Series D, doubling its valuation to $2.3 billion. New investor Ribbit Capital joined the round, alongside continued participation from Andreessen Horowitz (a16z), which has backed the company since its early days.
It Ran Its Own Mortgage Servicer Before Selling the Software
Valon’s business model is unusual. Rather than selling technology into the industry first, it built and ran its own full-scale, licensed national mortgage servicing business on its platform, ValonOS, before opening that operating system to others. Within six months of offering ValonOS to the industry, it signed more than $200 million in contracted annual recurring revenue, with major servicers including Rithm Capital’s Newrez, Carrington Mortgage Services, and ServiceMac signing on. One in six outstanding US mortgages is now under contract to run on ValonOS.
The Carrington relationship stands out in particular. In August, Carrington acquired Valon’s own servicing business — roughly 810,000 loans — and adopted ValonOS as its core servicing platform going forward; the price wasn’t disclosed. That deal effectively took Valon out of the business of servicing loans directly, leaving it purely as a software and AI vendor. Two of the top 10 US mortgage servicers, ServiceMac and Carrington, are now live on ValonOS. “Replacing core servicing technology is a significant decision, and not one ServiceMac took lightly,” said Rod Hatfield, COO of ServiceMac. “After more than 30 years in this industry, I know meaningful change requires thoughtful modernization. Valon’s capabilities support our efforts to help our teams work more effectively, manage risk, meet compliance requirements and deliver the service and accuracy our clients expect.”
ValonOS replaces a servicer’s fragmented systems with a single operating system for loan data, investor reporting, operational workflows, compliance logic, and money movement. On top of it, AI agents handle work across mortgage servicing — from answering homeowner emails to allocating payments and running escrow analyses. “The bottleneck for deploying AI agents into regulated industries is context, not intelligence,” said Linda Du, co-founder and president of Valon. “Mortgage servicing is a heavily regulated, edge-case-driven business, and agents need three things to be effective and safe: structured servicing data and context, decision traces behind workflows, and the ability to execute deterministic actions. We spent six years running a servicer and developed an ontology grounded in how mortgage servicing actually works.”
Over time, Valon plans to apply the same architecture to other regulated lending categories beyond mortgage, including auto, personal, and student lending.
Two Harvard Classmates Who Saw a Broken Industry From Wall Street
Wang and Du were computer science classmates at Harvard. Wang worked in Goldman Sachs’ aviation finance group before becoming a principal at Soros Fund Management focused on private credit, where he reviewed mortgage portfolios and saw servicers’ poor customer service, sloppy recordkeeping, and compliance gaps firsthand — the experience that led him to start Valon. Du worked as a software engineering intern at Google, an investment banking analyst at Goldman Sachs, a quantitative researcher at Two Sigma, and an analyst at Samlyn Capital before joining Wang to co-found the company in 2019.
Wang said the new funding “lets us bring [ValonOS], and the AI agents that run on it, to every servicer in the country.” Ribbit Capital founder Micky Malka said: “Andrew, Linda, Brian, and the Valon team are taking on a part of financial services that is badly in need of better technology. A mortgage is the biggest bill most families will ever have, yet the companies that service those loans still rely on complex, hard-to-navigate legacy software.” The “Brian” he refers to is Brian McGrath, an early board member since Valon’s founding who recently joined full-time as Executive Chair and Head of Engineering — not one of the company’s 2019 co-founders.
Andreessen Horowitz general partner Angela Strange said: “Servicing is the hardest, but also the stickiest, way to enter one of the largest debt markets in America. You have to turn regulation into code, get multiple licenses, and earn the trust of the biggest asset managers in the country. Valon has built the operating system for a $13 trillion mortgage market and is poised to do the same in other asset classes. We’ve backed Valon from the start, and have come back in every round since.” The $13 trillion figure refers to the total outstanding US mortgage debt, not annual servicing-industry revenue.
Valon started with a $3 million seed round in 2019, raised a $50 million Series A led by a16z in 2021, and a $43.9 million Series B the same year (backed partly by Starwood Capital) at a $590 million valuation. In 2024, it raised a $100 million Series C led by WestCap, pushing its valuation to roughly $1.1-1.15 billion — a figure that has now doubled to $2.3 billion with this Series D.
Competitors: A Legacy Giant and AI-Native Challengers
The dominant player in mortgage servicing technology is Black Knight‘s core servicing platform, MSP. Black Knight was acquired by exchange operator ICE (Intercontinental Exchange) in 2023, and MSP remains the industry’s standard legacy system — exactly the incumbent Valon is targeting.
Sagent, spun out of Fiserv and majority-owned by Warburg Pincus, competes head-on with its “DARA” platform, pitched as a modernized alternative to legacy MSP. Few competitors, however, share Valon’s approach of running its own servicing business before selling the software.
On the origination side, AI-focused players like Blend Labs and publicly traded Better.com operate in a different market — loan origination rather than servicing. In the broader category of deploying AI agents into regulated industries, startups like Norm AI (compliance automation for finance and pharma) and Hadrius (regulatory compliance for broker-dealers) share a similar thesis. Du’s framing — that the bottleneck is context, not intelligence — echoes a theme that recurs across this broader category of regulated-industry AI agent startups.
MORE FROM THE POST
- Co-Brand Card Platform Imprint Raises $130M at a $1.9B Valuation as AI Reshapes Commerce
- Spend Management Platform Payhawk Raises $280M Series C at $2B Valuation
- Model ML Raises $250M at $1.5B Valuation for Personal AI Agents in Finance
- Corporate Spend Platform Ramp Raises $1.85B at $60B Valuation
- Numeral Raises $100M Series C for AI-Powered Sales Tax Compliance
- Andreessen Horowitz
- Andrew Wang
- Fintech
- funding
- Linda Du
- mega funding
- mortgage servicing
- Overseas funding
- Ribbit Capital
- Series D
- Valon
- ValonOS
Share
Most Read
- 1
- 2
- 3
- 4
- 5


Leave a Reply