Figure and the Crypto Mullet Strategy

Plus! Billions; Intel; Negotiating the AI Transition; Price Discrimination; Sanctions

Byrne Hobart Byrne Hobart

In this issue:

audio-thumbnail
The Diff August 25th 2025
0:00
/841.822041

Figure and the Crypto Mullet Strategy

The S-1 for crypto-adjacent lender Figure is a fun time capsule. It's a reminder of the two poles of crypto bullishness:

  1. You can create arbitrarily complex tools for speculation, including decentralized lenders, automated market-makers, all kinds of crazy incentives around locking and unlocking supply, etc., or
  2. In most domains, if you want to store and retrieve structured information you just want a database. But on occasion, you actually want some kind of decentralized, permissionless tool that doesn't require you to trust third parties.

For a while, there was a sort of corporate politics arbitrage that went something like this: an executive running some kind of back-office cost center at a bank would think "our tech stack is ten years out of date, but things basically work, and I don't think I can get extra budget to migrate from Sybase to Postgres. My boss doesn't recognize either of those words. But, what if I said we were moving some processes onto the blockchain—a term my boss definitely recognizes—and then use Postgres as a more-likely-to-work-the-first-time backup option?" Talk about error rates, redundancy, delays from default approvers being in the wrong time zone, etc., and you put people to sleep. But "We could move it all to the blockchain" will get them excited, at least if crypto is a hot category.[1]
The S-1 actually calls this out at one point early on, listing two big problems in their industry: "Friction in Access and Extension of Consumer Credit, with Challenging and Antiquated Funding Markets" and "Limited Real-World Success of Blockchain Technology." It's kind of beautiful, in a way: flip the order and you have "blockchain technology is a solution in search of a problem," followed by "finally, a relevant problem"! It's an exaggeration in two ways, since crypto does solve several real-world problems—like "how do we make a better version of gold", but also lots of fiddly problems involving transferring money across national borders and quickly clearing transactions. There are real companies in this cycle, and one of the ways that you can tell that they're real is that they're somewhat disconnected from the hype.

And it's good to have a blockchain-flavored pitch, because otherwise some of the story is pretty boring. For example, consider the securities lending business. If you want to sell something short, you need to borrow it.[2] If you were to imagine how this works, in the year 2025, you'd probably assume that there's some sort of centralized clearing entity, that probably uses some kind of auction-based mechanism to figure out what market-rate borrowing cost is, and that this market in securities lending reaches its equilibrium because every willing lender lowers their reserve price until their assets are lent out. That's how Figure does it, as part of their Figure Exchange product:

There is no intermediary, and instead pricing is set every hour by a Dutch-style auction. Specifically, the interest rate is set by beginning at a predetermined maximum rate and successively lowering such rate until the aggregate principal amount of bids received equals or exceeds the offering amount sought by the borrower, with the lowest rate at which such amount is achieved becoming the interest rate applicable to all accepted bids. This is done for each collateral pool type, hourly.

That's only really notable because the way the equities industry does it is completely different; there's a consortium of big banks that run a platform called EquiLend, whose workings are opaque and expensive enough that a few years ago the banks settled a lawsuit over EquiLend's anticompetitive behavior for half a billion dollars. You, as an individual investor, might appreciate collecting a little extra income from shares you own that other people want to bet against. But you can only express that view very indirectly in the current system, because it's fundamentally descended from a phones-and-faxes approach that is itself descended from even older and more manual setups.[3] In Figure's case, this is actually not an incredibly promising development. Figure Exchange launched in August 2024, and "has facilitated the trading of over $1 billion in volume without meaningful marketing efforts, of which $299 million was in the six months ended June 30, 2025." So that's at least $700m in the last five months of 2024, compared to less than half that in the first six months of 2025. Elsewhere, they note that "For the six months ended June 30, 2025, we did not generate revenue from Figure Exchange..." So it's a shrinking, non-monetized platform—but the back-office functionality makes way more sense than what incumbents are doing. And, to be fair, exchanges are network effect-driven businesses, where a new one can lead to a big bang of asset shuffling after which things return to a more sedate pace for a while.

What they're trading on that exchange, and what they're generating most of their revenue from, is originating home equity lines of credit. This is a sort of reverse-retail business, where intermediaries like Figure are buying from lots of small counterparties and selling to a smaller set of large ones, making money on the markup between retail and wholesale price. This is a weird way for a borrower to think of things, because the borrower's interaction is that they apply for line of credit, get access to some amount of money, draw it down in chunks (for debt consolidation, funding a remodeling project, or whatever) and then pay it back over time once you reach the repayment period. But to a lender, looking at a portfolio of these, it's long-duration consumer credit with some collateral behind it, i.e. given a big enough sample size it's just another kind of fixed-income product. This leads to a division of labor where some companies specialize in origination, and then once they've made a loan they immediately flip it to someone who manages a fixed-income portfolio. (That "someone" does not have to be an active manager; it can be a structured product that's designed to vacuum up all of the HELOCs fitting certain criteria, and which has agreements with originators to buy them in bulk.) And even thinking of a set of originators and a different set of long-term lenders oversimplifies things, because there are other layers. Figure does some of its own direct marketing, but its big growth has come from white-labeling its product for other HELOC originators. So the division of labor is that somebody else is doing the late-night cable TV commercials, the billboards, the direct mail, etc., and then when that interaction gets a customer into the borrowing flow, Figure is the company making that transaction happen as quickly and cheaply as possible.

You can think about the various counterparty relationships here and imagine cases where "Blockchain!" is a magic word that makes things more efficient. But you can also imagine someone caring about the speed of various back-office processes, and know that in a regulated space you generally can trust third-parties because they'll be sued out of existence if they lie to you. Figure does have another business that's blockchain-adjacent: they'll lend against your crypto. But this is frankly less interesting than the high-throughput HELOC factory that uses, but doesn't need to use, a blockchain.

Longer-term, there can be strategic advantages to this, but they're downstream from more prosaic ones. If Figure's products end up on crypto rails, it's easier for crypto-native and fiat-naive counterparties to onboard than the reverse. So it does position them well for a world where more financial innovations are stablecoin-native rather than US dollar-native, and where they're first instantiated as a smart contract rather than as a legal contract. But that's not the kind of bet that you really want to use to underwrite an operating business that, again, produces GAAP profits by doing a good job of offering an existing financial product.

Figure does mention that there are other businesses they’d like to get into, that they currently can’t for regulatory reasons. But this leads to an incredibly salesly bit from the “risk factors” section of the S-1:

Due to our regulated status in several jurisdictions and our commitment to legal and regulatory compliance, we have not been able to offer many popular products and services, including products and services that our unregulated or less regulated competitors are able to offer to a group that includes many of our customers, or expand our offerings to new jurisdictions where we are currently not allowed to offer products, which may adversely impact our business, financial condition, and results of operations.

In other words: “Warning: the only way we’ll be able to access incredibly lucrative business opportunities that have already been proven to work is if the Trump administration somehow decides to deregulate crypto a bit, especially in a way that gives American companies a leg up. You should never assume that this ridiculously profitable outcome is inevitable.”

What's really interesting about Figure is that on the financing side, they're running a blockchain mullet strategy. When crypto is hot, as it is right now, they can say: we're a crypto company, we report GAAP profits, and our profits are not entirely from either betting that crypto prices go up or being the house for many other bettors. And, when that hype dies down, they can say: we're a HELOC originator whose streamlined process leads to loans getting made in a quarter of the usual time, and the bigger we get the more standard our level of service will be (and the more kinds of consumer credit products it will extend to). And if that story helps them raise money to bet on crypto when crypto's weak, and the crypto story funds R&D for HELOCs and other conventional forms of consumer credit when crypto's exciting, Figure's always in a position to make a contrarian bet, if they want to.

  1. Something similar is happening with generative AI, where partners at large accounting firms, CEOs of banks, insurance companies or pharmaceutical companies are, for the first time in 15 years, dying to know what their companies digital transformation roadmap looks like and whether it includes leveraging "proprietary internal data” in concert with LLMs to gain a competitive advantage. Any company who helps CIOs answer that question in the affirmative has seen a hyper-acceleration in the last year (see: Palantir). ↩︎

  2. You can get equivalent economic exposure to a short sale in other ways, like buying puts or being party to a total return swap, but in general those activities are a lot more economical if actual short-selling is possible. ↩︎

  3. In Beat the Market, Ed Thorp has an aside about this: "In practice, the loan clerk, who is responsible for locating securities, calls loan clerks in other brokerage houses. Depending on his stamina and persistence, he may call many loan clerks in search of the certificates. Usually he develops a relationship with some loan clerks, perhaps as few as two or three, and if they cannot accommodate him, he says the securities are unavailable for loan... Often an issue is “scarce” because the loan clerk lacks energy or contacts..." It's not that bad today, but if you use a service like Interactive Brokers, which offers an option to split the proceeds of lending your shares, you don't have control over whether or not your shares get lent, even if, for example, the borrow on something you own is 10% and you'd rather lend it out at 8% than not lend it at all. They'll either lend at the rate they set or not lend at all. ↩︎

SPONSORED## Elsewhere

You're on the free list for The Diff. Last week, paying subscribers read about Via and how to shift the efficient frontier between vehicle size and scheduling ($), the question of who will actually make big companies fully adopt AI tools ($), and when AI labs have an incentive to downplay AI investments' returns ($). Upgrade today for full access.

Upgrade Today

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

Find a Role

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.

Find Talent with Diff Jobs

Elsewhere

Billions

ETHZilla, formerly a shell of a biotech company and now an Ethereum treasury play, filed to sell up to $10bn in equity, which it will use for ETH. (To be honest, when I first saw this story I assumed it was a prank. Ten billion, really? But it might be audience segmentation instead.) But it's also a clever, exploitative move, in the sense that it acknowledges an obvious reality of crypto-treasury strategies: any company with a bit of cash can do one, and, if the premium is high enough, even companies without cash can pull it off by doing a PIPE to get their initial crypto stash. If that's the case, how do you differentiate? One bet is to just be the single biggest crypto treasury company for your particular token. If someone comes up with a bigger number, it's going to be a clear reference to ETHZilla, and that means they grab the spot as the default ETH treasury play rather than one of many competitors.

Intel

So, Intel has raised some money from the US government, in exchange for equity, and discovered that previous money sent to them by the government should have been reciprocated with equity all along. It's actually an incredibly tempting approach to couple corporate subsidies with equity ownership. If the government is making a company better-off, it seems only fair to give the government a stake in the upside, perhaps at a valuation that still makes it an obviously good deal for the company.

The problem is that the long-term incentive is for the company to arrange itself around needing constant infusions of capital. The more Intel raises this way, the more attractive further subsidies are, since they help bail out the previous investment. If the government is going to take an equity stake in a previously-private company, but not take it over completely, the only structure that aligns incentives correctly is for them to be straitjacketed into only making the investment one time, and committing to sell it down in the future.

Negotiating the AI Transition

Netflix has released guidelines to creators on their use of AI, where at least one of the rules explicitly protect existing jobs ("GenAI is not used to replace or generate new talent performances or union-covered work without consent."). It's hard to imagine something that is both useful and also doesn't replicate the work of any talent or unionized professional, so part of what it's doing is shifting blame. Netflix can't stop creators from using AI, at least in some cases, but they want to make sure it was the creator's decision, not theirs, almost like an AI evolution of section 230.

Price Discrimination

The first fees for alternative asset managers were set somewhat arbitrarily—Alfred Winslow Jones,founder of the first hedge fund, had spent the first half of his life in a non-financial milieu (getting a sociology PhD at Columbia, hanging out with Hemmingway, etc.), and he claimed that his fee structure was inspired by Phoenician merchants, at a time when hedge fund investors had less immunity to show-offy IQ signaling than they do today. And that fee model was ported over to private equity and venture.

Hedge fund fees have come down since then, though some versions of the hedge fund model let funds structure their strategy around justifying their fees rather than setting fees that match the returns from a strategy they've chosen. PE doesn't have that flexibility, so PE firms have had to start offering fee discounts ($, FT). In a way, this is yet another instance of the convergence between public and private equity. If investors are looking at the same pool of companies, and if PE's target capital structure isn't all that different from what a mature public company would optimize for, fees can start to really eat into returns.

Sanctions

The way sanctions typically work is that the real bite comes from the secondary sanctions, where banks have to choose whether they do business with a given sanctioned company, or with the US dollar system, because they won't be allowed to do both. But that's not entirely true; it would be too disruptive to apply these sanctions to big Chinese banks, so Russia still has some access to the global financial system through China. This still leads to lots of inconvenient frictions, and presumably Chinese counterparties have the negotiating leverage to force their Russian business partners to absorb most or all of the costs. And there's still a possibility that the Trump administration will go ahead and do something dramatic and disruptive like applying sanctions to a major Chinese bank over ties to Russia, regardless of how disruptive that might be. Applying sanctions starts out as an easy decision, but after that it keeps leading to harder ones.