How Valon Is Rebuilding Mortgage Servicing for the AI Era
a16z Deep Dives · 2026-10-06 · 39 min
https://www.youtube.com/watch?v=ZBDGuT6EGPYVideo summary
Valon rebuilt mortgage servicing for a $13 trillion market, reaching 3× efficiency and preparing to transfer 4 million loans with AI.
Valon co-founders Andrew Wang and Linda Du tell a16z’s Angela Strange why they rebuilt mortgage servicing from scratch in a $13 trillion industry still running on pre-internet systems. Andrew traced the idea to reconciliation problems he encountered at Soros; Linda described how legacy records can leave families without vital loan history and make rising escrow costs harder to manage. Valon spent years navigating state licensing and translating federal rules and 50 states’ regulations into software, then proved the platform on its own servicing business. The founders say it is about three times as efficient as traditional servicing, with 70–80% operating margins, and that customers signed more than $200 million in deals within six months of the software launch. They see AI enabling agents to handle complex, tailored workflows while continuously evaluating results. After selling its servicing business to Carrington, Valon is focused on its software platform, including a planned transfer of 4 million loans from Rhythm. They argue the same regulated, money-moving infrastructure could extend beyond mortgages into healthcare and other sectors.
Chapters
- 0:00Intro: Mortgage is a top-three undisrupted industry still behind 2000s technology
- 1:02From Soros to mortgage math: Andrew Wang’s reconciliation failures inspired better infrastructure
- 3:0013 trillion still on 1960s systems: Legacy records and escrow failures harm homeowners
- 9:38Why build the whole stack from scratch: Valon chose a servicer that grew from one loan
- 11:47The 3–5 year licensing gauntlet: Non-QM loans and a three-year New York approval
- 15:50Turning 50 states of regulation into code: Valon mapped federal rules and tested its framework
- 18:00Winning the first customers: Three-times efficiency and Valon’s first-customer flywheel
- 22:34What AI unlocks now: agent-led disaster outreach, loan experiments, and voice AI
- 29:47Everything is servicing: commercial tenant data, healthcare claims, and mortgage relationships
- 32:52Rhythm's 4M-loan transfer: $200M in deals, Newrez proof, and enterprise change management
This is a Tier 1 public summary
Whether the chapter key points, section summaries and mind map are public is up to the person who shared it. Want the full analysis?Submit one yourself.
More from this channel
To Regulate AI Effectively, Focus on How It’s Useda16z Deep DivesMartin Casado argues AI laws should target harmful uses, while regulatory uncertainty pushes U.S. startups toward Chinese open-source models.
Inferact: Building the Infrastructure That Runs Modern AIa16z Deep DivesInferact founders explain how vLLM serves 400,000–500,000 GPUs and why they’re building an open-source inference layer.
How Palantir Scaled: Why the Best Software Is Built Backwardsa16z Deep DivesPalantir Chief Architect Akshay Krishnaswamy explains how forward-deployed engineers turn customer pain into reusable products.
Mintlify and the Transition From Human Docs to Agent Infrastructurea16z Deep DivesMintlify cofounder Han Wang says docs are becoming AI infrastructure, with 20 million people visiting Mintlify-powered sites last month.
Braintrust CEO on Where Engineering Actually Matters in AIa16z Deep DivesBraintrust CEO Ankur Goyal argues engineering should focus on evals and harnesses, not models, as Chinese models approach 95% on a benchmark.
Related analyses
Building Cyber Defense for the Agentic Eraa16zArmadin CEO Kevin Mandia says AI attacks operate at machine speed, and Armadin has found over 90 production zero-days since January 2026.
Why AI Is Reinventing How Businesses Buy Everythinga16zLio CEO Vladimir Keil says AI agents can manage procurement end to end, where one missed aircraft-part email could cost hundreds of millions.
Why Specialized AI Could Beat The God Modela16zOpenRouter’s Alex Atallah and Replit’s Amjad Masad argue specialized AI can outperform a single god model, with fusion reaching frontier quality at 40–50% of the cost.
The Current State of Consumer AIa16za16z finds only 4.5% of US consumers pay for AI, while OpenAI’s ads reach a $1 billion annual run rate.
EP157 - AI Agent 大戰全面開打！深度解析 OpenAI Dots、Muse、Grok Bot、Instinct科技浪Instinct、Meta Muse、Grok Bot、OpenAI Dots 帶動個人 AI Agent 浪潮，CPU 需求也隨雲端 VM 擴大。