Altisource: 26% Growth, a 24% Drawdown, and a $45 Million Promise
A growth story the tape refuses to underwrite
On paper, Altisource Portfolio Solutions' third quarter reads like a breakout. Service revenue rose 26% to $49.8 million, adjusted EBITDA climbed 17% to $4.3 million, adjusted pre-tax income jumped 59% to $2.1 million, and adjusted EPS rose 30% to $0.13. Yet this roughly $61 million micro-cap entered the report down 24% over 90 days, 28.8% below its June 30 high of $7.50, and 97% below where it traded in 2010. That gap — strong topline, dead tape — is the whole story. The market is not paying for growth it suspects is expensive and low-quality, and management spent this call arguing the mix is a down payment on a much larger earnings base.
The engine of the quarter was customer wins, the company's top-ranked theme and a genuinely new center of gravity versus prior quarters dominated by delinquency and foreclosure statistics. Wins secured since 2024 generated $13 million of revenue in the quarter — $51.9 million annualized — and the servicer and real estate segment added an estimated $2.7 million of stabilized annual wins, with a $12.8 million weighted pipeline behind it. Bill Shepro framed it plainly: “Service revenue of $49.8 million increased 26% over the same quarter last year, and adjusted EBITDA of $4.3 million increased 17% over the same period.” — Bill Shepro, CEO · 2026-10-29
The real structural change: dependency is melting
The most underappreciated line in the release is the diversification stat. Revenue from customers other than Onity, Rithm and their portfolio companies grew 96% year over year and now represents 72% of service revenue, up from 46% a year ago and 65% last quarter. For a company that spent 2025 and 2026 on prior calls talking about loss of revenue and dependence on Onity, that is the pivot: the concentration problem is being solved by outgrowing it rather than waiting for it to heal. Total revenue has now grown year over year for consecutive quarters after bottoming near $34 million in 2023, a shape that only recently turned up.
Beneath that sits a quiet leading indicator. Hubzu inventory within Marketplace more than quadrupled to 26,000 assets from 5,700 a year earlier, and rose 16% sequentially. Because inventory takes nine to twelve months to season into sales — a timeline management spelled out on the prior call — today's stock is tomorrow's transaction volume. “Typically, when you receive an REO file, it could take anywhere from, let's say, 9 to 12 months to sell.” — William Shepro, Chief Executive Officer · 2026-07-23 The Default market remains below pre-pandemic norms, but seriously delinquent loans and foreclosures totaled 872,000 in August, up 26% year over year. The macro wave is building; Altisource is positioning itself to be paid before it breaks.
Why the stock won't pay up: margins and cash
Here is the tension. Adjusted EBITDA margins of 8.5% declined from a year ago, and the analyst on the call — lone questioner Timothy D'Agostino — opened by pressing on gross margin slipping from 27.5% in Q1 to 23.4% in Q3. Gross margin sits at roughly 25%, down nearly 5 points year over year. The CFO's explanation was mix, not decay: “we have had much higher growth in the origination segment than we have experienced in the servicer and real estate segment.” — Michelle Esterman, CFO · 2026-10-29 The bet is that today's lower-margin revenue is an investment in tomorrow's operating leverage.
The first would be in the servicer and real estate segment... we were already incurring the cost to support growing revenue from higher inventory and referral volumes, but that revenue we are not going to incur more costs as that revenue grows.
That is a credible mechanism — but it is still a promise, not a delivered margin. Meanwhile the balance sheet is the elephant in the room: effective net cash is roughly -$144 million, and free cash flow less stock comp was -$8 million last reported quarter. Altisource bought back only $900,000 of its term loan in the quarter. With receivables running at 53% of revenue, growth itself consumes cash. This is why the market discounts 26% topline growth: it arrives with negative free cash flow and a leveraged capital structure.
Project 45: the $45 million moonshot
The call's new organizing principle is Project 45 — reaching $45 million of run-rate adjusted EBITDA by Q4 2028. Against roughly $17 million annualized today, that is a near-tripling, delivered on the core building blocks of ramping wins, operating efficiency and flat corporate costs. The efficiency lever is explicitly an AI automation story — used to shorten software development cycles and streamline Marketplace, field services and valuations. It is a quintessentially 2026 pitch: an AI narrative bolted onto mortgage default servicing, a far cry from the AI-data-center tape that has been fading in the broader market.
Two quieter signals matter. First, coverage is thinning — only one analyst asked questions this quarter, versus two on the July call, and the prior liquidity questioner Shachar Minkove was absent. Second, the language of the call has rotated decisively from macro-waiting to self-help: Origination markets were described as difficult, with unit volume down 6%, yet the segment grew 65% — market share, not market lift. That is the real argument. But as long as margins compress and cash burns, the equity will keep discounting the promise. The next two quarters either validate Project 45 or expose it — and the Hubzu inventory already sitting on the books gives the bulls their best evidence that the revenue half of the equation is real.