Open in interactive viewer → charts, metric popovers & call review

Figure Is Buying Back the Story It Won't Fully Explain

A 107%-growth lender at a 58% EBITDA margin trades 65% below its January peak — and just authorized a buyback worth ~9% of its market cap while deferring its new margin metric for a second straight quarter
FIGR · Earnings Call · 2026-11-12

Figure Technology Solutions reported on November 12 into a tape that has already made up its mind. The stock sits 65.3% below its January 16 peak of $73.91, down 21.7% over the last 90 days alone and 37.6% off its late-August high of $41.14. Against a $4.49 billion market cap, the company used the same morning to authorize a $400 million repurchase — roughly a ninth of itself — with “We have been trading below our IPO price recently despite over 100% annual volume growth since IPO.” — Michael Tannenbaum, Chief Executive Officer · 2026-11-12

That is the central tension of this report: an extraordinarily fast-growing, currently profitable lender whose equity is being priced as something else entirely. And there is a plausible reason for the gap — Figure is not one story, it is three, and only one of them is currently earning the multiple.

The Kiavi Math: More Volume, Less Revenue Per Dollar

The Kiavi acquisition closed on September 1 and delivered roughly $670 million of volume in a single month. The loan marketplace printed $5.8 billion for the quarter (up 135% as reported), $5.1 billion of it organically (up 107%). The 58% adjusted EBITDA margin came on $164 million of adjusted EBITDA, up 91% year-over-year. Net take rate was 3.5% including one month of Kiavi and 3.7% stand-alone — better than the guidance management gave in August.

But Kiavi's products are short-duration. Residential transition loans ran roughly $600 million of September volume; DSCR loans ran about $60 million. And short duration means short fee duration. The CFO was blunt about the consequence:

We expect gross take rate from shorter duration RTLs to be in the 1.5% to 2% range, whereas for longer duration Connect HELOC, we expect to be in the 2.9% to 3.4% range over the next 6 to 12 months. ... That's why we're shifting our focus to gross take rate going forward rather than net take rate.

Minchung Kgil, Chief Financial Officer · 2026-11-12

So the metric is changing precisely as the mix is worsening it. Q4 combined gross take rate is guided to 3.1%–3.4% on $6.2–$6.7 billion of volume. Management's defense is that this is a capital light trade — Connect volume carries no customer acquisition cost, fees are tiered rather than negotiated, and the entry of a large whale simply pushes a partner into a lower price band. “Our pricing doesn't reflect negotiated partner rates. So when we talk about whales and pricing, what we're talking about is an expectation that we hit high volume tiers.” — Michael Tannenbaum, Chief Executive Officer · 2026-11-12

The problem is that Contribution margin — the metric that is supposed to replace take rate as the real scoreboard — still hasn't been defined publicly. In February the CFO put it around 80%. In August, the CEO promised it "next earnings call": “We just want to give it when we have full line of sight into Kiavi to make sure that we really only have to explain it one time to you all and not waste your time.” — Michael Tannenbaum, Chief Executive Officer · 2026-08-13 Asked again in November: “We will talk more about the details of how contribution margin comes together in our next earnings call.” — Minchung Kgil, Chief Financial Officer · 2026-11-12 Two deferrals is a pattern, not a schedule. Investors are being asked to accept that a 2% gross-take-rate dollar is worth a 6% balance-sheet dollar, with the evidence promised later.

The reported numbers do at least support the profitability claim. Total revenue reached $226M in the latest filed quarter, up 113% year-over-year, with an operating margin of 36.8%. What the income statement does not show is the liquidity drag underneath: free cash flow, less stock comp, was −$61M. That is largely the plumbing of a marketplace — loans sit on balance sheet for weeks before sale — but it is also why a company earning $89 million of net income still leans on $1.59 billion of cash, $600 million of new unsecured debt and a $465 million revolver.

Navra: The Question Nobody Asked

The genuinely new item this quarter was not financial. Executive Chairman and co-founder Mike Cagney launched a separate wallet company, Navra, and Figure holds a 10% stake. He devoted a prepared-remarks section to it — an unusual allocation of airtime for an entity that isn't consolidated — and the framing was defensive:

The second company wasn't on their bingo card after the IPO, but Navra widens the market for what Figure builds, and that's why I'm spending the extra time to put the solution into the marketplace.

Michael Cagney, Executive Chairman and Co-founder · 2026-11-12

The logic is coherent: DeFi funding at SOFR+120 versus SOFR+250 through a warehouse line on roughly $600 million of HELOCs, with a $6.5 billion stablecoin borrowing pool against upwards of $2.5 billion a month of assets Figure puts on chain. The constraint is market size, not economics, and Navra is an attempt to widen it. But the absence of a single analyst question on founder attention allocation — in a stock down 65% — tells you the sell side has decided this is not yet a governance issue. It may become one.

The same build-out explains why Figure keeps showing up in an unrelated tape. The global real world asset basket has been a 30-day decliner, and FIGR is literally in that basket alongside exchanges and crypto miners. Figure is being traded as tokenization beta, not as a mortgage-fintech compounder — which is why the rate environment argument, that rising rates actually help a B2B HELOC shop, falls on deaf ears.

Two Things That Quietly Disappeared

The keyword record is revealing. Democratized Prime was Figure's single strongest keyword a year ago; this quarter it was its sharpest decliner. The reason is not strategy — it is a rename. “Prime, which was previously known as Democratized Prime, ended the quarter with lender supply of $825 million.” — Minchung Kgil, Chief Financial Officer · 2026-11-12 Likewise capital light marketplace language has given way to "balance sheet-light." Vocabulary churn, not business churn — but worth flagging when the sales pitch is itself in flux.

What is intact is the capital-markets engine. Q3 saw $2.7 billion of loan sale activity and eight securitizations, with investor demand broadened by marquee AAA buyers and new foreign accounts. Management argues there is "no friction" pivoting between whole-loan and securitized execution. That is the moat the buyback is implicitly underwriting, and it is the one thing the quarter genuinely validated — along with cost synergy work, where $30 million of the $35 million target was actioned at close.

So: a company growing volume at triple digits, earning $89 million in a quarter, buying back ~9% of its shares, and asking the market to grade it on a metric it has twice postponed defining. The Sierra agent — 29% more funded dollars per approval, ~$260 million funded through it — is the most under-discussed asset in the deck, and the most likely reason the story eventually gets re-rated. But the immediate question is narrower: does a 3.1% gross take rate on $6.5 billion of Q4 volume produce enough contribution dollars to justify the multiple management just put on its own stock? Until Figure shows that math, the buyback is doing the talking.