Join our mailing list

Rebellionaire sits down with Paul Gu — Upstart co-founder, AI architect, and incoming CEO — to talk model upgrades, product expansion, insider buying, and why the best is still ahead. The Man Behind the Models Paul Gu doesn't do things for optics. That's one of the first things you pick up when you talk to him. He co-founded Upstart in 2012 alongside Dave Girouard, and for the past 14 years he's been the person most responsible for the AI and machine learning that makes the w

Upstart grew originations 86% in 2025, flipped to profitability, crushed the balance sheet concern, and just handed the keys to co-founder Paul Gu. The stock dropped 25%. Here's why the fundamentals have never been stronger — and why the selloff might be creating a massive asymmetric opportunity. The Quarter That Was Supposed to Break the Thesis Heading into Q4, Upstart had every chance to disappoint. After a confusing Q3 that involved a macro-driven tightening credit box, a

Upstart’s stock is still being judged like it’s 2022. The business isn’t. That disconnect is the entire setup going into Q4 earnings. This article summarizes original Rebellionaire research based on proprietary indicators and first-party data analysis. What is this report about? This report analyzes Upstart’s Q4 2025 earnings with a focus on whether the company is still a balance-sheet lender or in an intentional AI incubation phase. It examines funding structure, model behav

If you’ve been hanging around Rebellionaire for a while, you know Henry’s Upstart Trustpilot Index (UTPI) has turned into one of our favorite “weird but useful” data sets. It’s not GAAP. It’s not official guidance. It’s people leaving reviews on the internet—and yet, quarter after quarter, it has mapped surprisingly well to what Upstart actually reports. Henry just dropped the latest UTPI + UpstartIQ update (you’ll find the full PDF embedded at the bottom of this post), and t

UpstartIQ used to treat revenue like one big bucket. Everything got blended together, which made the model simple… but not very accurate. A lot has changed since then. Now the framework splits revenue into two distinct categories that behave very differently: 1. Non-Core Revenue — Where Things Get More Complex This is the portion of the business driven by balance-sheet mechanics and servicing economics, not borrower flow. Upstart now models: Net interest income  as a percenta

Editor’s note for investors/builders:  Upstart just delivered one of those quarters that looks messy at first pass and more interesting the closer you get. What actually happened? Upstart posted $277M in revenue (+71% YoY), an 11% GAAP net margin, and 26% adj. EBITDA margin. That’s a healthy P&L—just shy of their own revenue guide. The miss wasn’t demand; applications hit a three-year high (2M+). It was the model tapping the brake: approvals came in tighter, average loan size

Q2 proved the model isn’t dead (102% YoY revenue growth, swing to GAAP profit). The stock still got crushed on balance-sheet optics ($600M convert, R&D loans for new products) and a later-corrected delinquency scare. Heading into Q3 (Nov 4), the tape is hypersensitive to funding durability  and how fast credit risk moves off balance sheet . (See BTIG correction and UMI context in the charts on pp. 8–10.) UPST Q3 Earnings Preview What changed under the hood Mix is diversifying

Download the in-depth report 👇 After a late-August wobble, UTPI’s 28-day moving average ripped from 20.1 to 31.7—up 57.7%  in just over...

TL;DR:  Upstart’s Trustpilot Index (UTPI) rolled over hard in early August. Hitting the Q3 review target now requires a late-quarter...