Winning and Losing With Tokenized Deposits

3 news stories, 2 reading recommendations, & 1 question.

In partnership with

Happy Monday, Fintech Takers!

I trust your weekend was wonderful and I hope that your college football team (if you have one that you root for) won.

My kids started school today and I was reminded (as I always am this time of year) what a magnificent thing public education is. You walk with little humans holding your hand to a building that is filled with some of the kindest and most dedicated professionals in our society — people who are 100% in it for the right reasons — and your little humans go off and learn everything they’re going to need to know for the rest of their lives, while also having a ridiculous amount of fun.

Modern state-sponsored public education has existed for roughly 200 years, which means that you and I are in the incredibly lucky 4% of humans, since the start of recorded human history, to have had free access to a service that our predecessors would have done anything to get their hands on.

We hit the lottery, basically.

- Alex

P.S. — Please sign up to attend this upcoming virtual event. If for no other reason than the title, which I am immensely proud of!

Was this email forwarded to you?

Subscribe Here

Sponsored by Middesk

41% of businesses will give you 24 hours to approve them. After that, they apply somewhere else.

Now that payments settle in seconds and purchases take one click, business onboarding is being held to the same standards.

A multi-day onboarding queue reads like a choice, so applicants treat it like one.

Middesk customers are showing that faster onboarding doesn’t require fewer checks.

Every applicant is still automatically verified on identity, ownership, web presence, and risk signals.

When those signals tell a consistent story, the business is approved. When they don't, a human steps in.

The manual work gets reserved for the applications that earn it, which is how teams onboard more businesses without hiring their way there.

Get the case study

Cho senshu by Kamisaka Sekka (1904)

3 FINTECH NEWS STORIES

#1: Winning and Losing With Tokenized Deposits

What happened?

39 state bank trade associations are teaming up on blockchain and tokenization:

Thirty-nine U.S. state bankers associations have formed BankChain Alliance, formalizing a bank-industry effort to build a common blockchain network for tokenized deposits, stablecoins, smart payments and automated settlement.

The alliance is targeting a 2027 launch but is not yet an operating payments network. Its Aug. 25 announcement said it is still selecting a technology partner and plans to make the network interoperable with other systems while inviting banks across the country to become owners.

BankChain describes the project as industry-owned, designed and governed. That structure is intended to give community and regional banks a stake in infrastructure that might otherwise be controlled by large banks, core-technology vendors or crypto companies.

The Financial Accounting Standards Board (FASB) proposed guidance on when companies may classify stablecoins as cash equivalents under generally accepted accounting principles:

Three conditions apply. A qualifying digital asset needs an on-demand contractual redemption right, a direct redemption right with the issuer for a known cash amount, and segregated reserves held at no less than a one-to-one ratio in short-term, highly liquid assets.

The board explicitly rejected a looser standard. FASB said secondary-market liquidity alone would not be sufficient, meaning a holder must have a direct claim on the issuer rather than merely the ability to sell the token on an exchange.

And two Dallas Fed economists estimated the impact that tokenized deposits will have on banks’ businesses:

Tokenized deposits could reduce U.S. banks’ capacity to hold long-term interest-rate risk by about $700 billion if they make depositors 10% more sensitive to rates.

A separate scenario found that if tokenization causes deposits to leave banks 10% sooner, banks could lose about $580 billion of capacity to absorb the interest-rate risk of long-term loans and securities, the economists estimated.

So what?

My very basic mental framework for stablecoins vs. tokenized deposits is that while they are both tokenized forms of money, tokenized deposits are generally better for banks (they can pay yield, they can be lent against, they map cleanly onto existing compliance requirements, etc.) and stablecoins are generally better for end customers (more interoperable, natively cross-border, more control through self-custody, etc.)

Now, there are things you can do to make tokenized deposits more attractive to end customers, compared to stablecoins. The GENIUS Act’s prohibition on payment stablecoins offering yield was a (shoddy) attempt by Congress to artificially hobble the utility of stablecoins, relative to bank deposits (tokenized and otherwise).

FASB — an independent standard-setting body empowered by the SEC — is putting another compliance feather in tokenized deposits’ cap. Its proposed guidance for classifying stablecoins as cash equivalents includes a few different tests. The one that really stings for stablecoin issuers is the requirement that the holder of the stablecoin have a direct redemption right with the issuer for a known cash amount. This is not the outcome that large stablecoin issuers like Circle wanted to see. Circle’s distribution is split between direct (some large corporates have Circle Mint accounts, which allow them to mint and redeem USDC directly) and indirect ( through partners like Coinbase). Today, I would guess that a large percentage of indirect distribution is for consumers and small businesses (who don’t really care about how stablecoins are treated from an accounting perspective), while those large corporates that care about the accounting benefits probably mint and redeem with Circle directly. However, Circle’s future growth hinges heavily on indirect distribution through partners and the payments infrastructure that the company is building, and FASB’s proposed requirement for direct redemption will (if it stands) limit the company’s ability to grow on the corporate side through these channels (which is why the company opposed this specific requirement in its comment letter to FASB). Tokenized deposits, which are literally a direct customer claim on a bank’s balance sheet, sail right through this FASB requirement, which will (if it stands) strengthen the case for tokenized deposits over stablecoins.   

Beyond regulatory compliance, banks can also collaborate to make tokenized deposits’ lack of native interoperability less of a competitive disadvantage. That’s what the Clearing House (a payments-focused consortium owned by the big banks) is doing with its tokenized deposit network. And it’s what state bank trade associations — whose member banks hate and fear the big banks and their consortiums — are doing with BankChain.

The details are a bit sparse on BankChain, but a few initial thoughts from me:

• The bank trade associations aren’t committing to a specific product structure yet. They are reportedly considering everything, across tokenized deposits, stablecoins, smart payments and automated settlement. However, I will be surprised if the first and primary product isn’t a tokenized deposit offering.

• Participation by the state associations does not guarantee participation by the individual banks that belong to those associations. They have some serious work to do to actually get banks signed up. A 2027 launch may be overly optimistic.

• They apparently haven’t picked a technology partner for this initiative yet, but according to reporting from American Banker, the group wants an ownership stake in the technology partner it selects. That’s interesting and makes me think it’s more likely to partner with an existing bank-focused network like Cari or Hazel, rather than a bigger infrastructure provider like Stripe (though Stripe would be very smart to aggressively pursue this opportunity IMHO).

The irony in all of this is that, in their rush to avoid the fate of the incumbent in Clayton Christensen’s most famous drama, banks may still end up worse off than they were before. The Dallas Fed research, while speculative, paints a grim picture for banks, even if they manage to turn the market away from stablecoins and towards tokenized deposits.

The authors argue that tokenization weakens the frictions that make deposits sticky. The combination of smart contracts and agentic AI could move money to higher yields with no action from the holder. And, if that happens and deposits become even a little bit more rate-sensitive or leave a little bit earlier than they otherwise would have, banks’ ability to engage in maturity transformation will be severely hampered, resulting in anywhere from $580B - $700B (depending on the scenario) in reduced lending capacity.

The authors cite an interesting precedent for this possibility: A 2025 study of Brazil’s instant payment network Pix found that heavier usage of the system increased banks’ holdings of liquid assets, particularly government bonds, while reducing credit intermediation.

#2: The Unreliability of SPCPs

What happened?

The FDIC, NCUA, OCC, CFPB, HUD, DOJ, and FHFA (but not the Fed … notably!) have rescinded a Biden-era policy statement regarding fair lending:

[The] federal agencies have withdrawn a 2022 policy statement that encouraged banks and other creditors to use special purpose credit programs [SPCPs] to expand access to financing for underserved groups. The rescission took effect August 25, 2026, and affects guidance involving the Equal Credit Opportunity Act, commonly called ECOA, and its implementing rule, Regulation B.

The change does not eliminate special purpose credit programs altogether. Instead, it removes the agencies’ 2022 interagency statement and comes after a separate 2026 change to Regulation B that narrowed how certain characteristics can be used in these programs.

So what?

This is really just the formalization of a change that the CFPB made to Reg B earlier this year, when it prohibited for-profit creditors from using race, color, national origin, or sex (or any combination) as the common eligibility characteristic in an SPCP and made the remaining characteristics (religion, marital status, age, receipt of public-assistance income) essentially unworkable for for-profit creditors to use at scale.

Prior to encouragement from Biden-era regulators in 2022, lenders (especially for-profit lenders) hadn’t seriously explored the use of SPCPs, even though they were (and continue to be) permissible under ECOA/Reg B. My sense is that many of the big banks (Wells Fargo, Bank of America, JPMorgan Chase, U.S. Bank, etc.) started these programs in the 2020s more out of a desire to please regulators (and earn good press and CRA credits) than out of a strong belief that it was the most profitable course for the business. And yet, in the few years when these for-profit programs were running, I heard multiple anecdotes from the bankers who ran them that went something like this, “We did this because regulators were encouraging us to, but it’s actually been really profitable for us.”

These are just anecdotes, obviously, but I think there’s something to them. There are structural inequalities in our society that have produced gaps in different protected classes’ credit scores (race is the most obvious one). These gaps result in less credit being extended to those groups with lower scores. However, the theory behind the 2022 policy statement was that if banks could just be encouraged to take a bit more risk (in both their acquisition and underwriting) in working with these groups, they might discover that they perform significantly better than their scores might suggest.

This theory didn’t get much time to get tested before the second Trump administration slammed the window shut (no bank or fintech company is going to dare trying to stand up an SPCP right now). However, another practice that received encouragement from regulators (under both the first Trump administration and the Biden administration) seems likely to outlast this fair lending fracas. Cash flow underwriting, which reduces lenders’ dependency on credit scores, has already proven to be enormously popular with fintech and banking lenders, especially when they are prospecting in near-prime, subprime, and credit invisible segments.

Moving forward, my guess is that regulatory guidance on SPCPs will end up being so polarized that lenders will stop taking it seriously, one way or the other, and will, instead, focus on investing in technology that makes it easier to safely approve more applicants regardless of what the regulators say.

That’s what I would do, at any rate.

#3: Socure Secures More Money

What happened?

Socure raised some capital and bought a company:

Identity verification startup Socure raised $156 million in a series E extension round that valued the company at $5.2 billion …

The extension comes nearly five years after the company raised $450 million as part of its Series E round at a $4.5 billion valuation.

The latest investment included primary funding and a secondary tender offer for employees. It was led by Summit Partners, ⁠with participation from Goldman Sachs Alternatives, Wells Fargo, DocuSign and other investors.

Socure also announced the acquisition of Fravity, an agentic operations platform that automates fraud, risk and compliance work. Financial terms of the acquisition were not disclosed.

So what?

Some quick notes on this one:

• It’s roughly a 16% increase in valuation, which is probably why it’s being presented as an extension on the Series E rather than a Series F, even though Summit is a new investor in the company. Summit is not like the investors that gave Socure a staggering $450M in 2021. Those VC firms (Accel, Tiger Global Management, T. Rowe Price) were well-known (especially in 2021) for chasing hypergrowth startups at any cost, often without significant due diligence, which is why many of the valuations from their 2021 cohorts look nutty now. Summit is known as a more disciplined late-stage investor that looks for durable, metrics-driven compounders. The fact that it liked Socure enough to invest at even a marginally better valuation is a huge sign of confidence in what Socure has built since 2021.

• To put some numbers on that, the company reports that it’s profitable, that it’s currently seeing $364M of total ARR, growing 63% YoY, and has a net dollar retention of 133%.

• From a product perspective, the company’s journey has been interesting. It started off in identity verification for KYC during new customer onboarding. Its breakout solution — a synthetic identity score built on top of a large give-to-get data consortium — was a great example of what you frequently see in fraud, when a great product meets a growing risk vector and adoption just explodes. After that, the company has added to its product suite through acquisitions, buying Berbix (document verification) in 2023, Effectiv (orchestration and decisioning) in 2024, Qlarifi (BNPL credit reporting) in 2025, and now Fravity. Of those acquisitions, Effectiv was the most significant as it helped the company transition from being a data and scoring provider (which is a very precarious spot in the fraud solution stack to be in) to a workflow orchestration provider (which is a much more stable position). I wonder to what degree Fravity’s agentic technology will end up being infused into Socure’s core workflow and decisioning capabilities, versus continuing to be used solely as a tool to speed up manual investigations.

• Despite this latest fundraise, Socure isn’t entirely out of the range of an acquisition, if growth slows down. The credit bureaus probably couldn’t swing it, but someone like a Fiserv (if it was healthy and back on track) might be able to, and one of the card networks certainly could, if they took a liking to Socure’s product suite and book of business. However, all signs at the moment point to an IPO as the most likely exit, though timing for IPOs is always a tricky needle to thread.

2 READING RECOMMENDATIONS

#1:  A market still needs two sides (by Kristen Anderson, Skeptical Optimism) 📚

Reading Kristen’s first newsletter (built on top of the Finity Newsletter Platform!) is a breath of fresh air. Insightful, reasonable, precise, and filled with amusing turns of phrase and analogies to parenting young children.

Read it! And sign up as a subscriber to Skeptical Optimism!

#2: It’s worse (by Joshua Gans, Mess and Magic) 📚

I read the Dwarkesh article on the Hugging Face incident and found it to be well-written but (as others have pointed out) strangely anthropomorphized. I don’t have a problem with anthropomorphism in AI generally (it can be helpful!) but I felt like the Dwarkesh piece was unnecessarily dramatic.

Still though. The Hugging Face incident is really weird! It seems to involve a level of coordination, deception, and (at times) desperation that has shocked AI and cybersecurity researchers.

And when those folks are shocked, we should probably take time to examine and update our prior beliefs, as this very good article by Joshua Gans argues.

*Bonus: Will AI's Next USD 500 Billion Be Serviced In Excel? (by Stitch) 📚

Earlier this month, NVIDIA and six of the world's largest asset managers agreed to bankroll $500+ billion in AI infrastructure, the same way banks finance airplanes: lend against the hardware, get repaid by what it earns. Stitch's new piece asks whether the software running that servicing today can keep up. Worth your read.

*This rec is brought to you by one of our fantastic brand partners.

1 QUESTION FROM FINITY

There are a TON of interesting questions being asked in Finity (our digital community for fintech and banking nerds). I’ll share one question, sourced from the community, each week. However, if you’d like to join the conversation, please apply to join!

How will industries that depend on a high level of customer trust (banking, healthcare, etc.) need to change if AI fundamentally alters the safety of computer systems?

This Twitter thread by Will Manidis pokes at this question in an interesting (if perhaps overreacting) way and I’d be curious to get y’all’s thoughts on it!

If you have any thoughts on this question, reply to this email or DM me in Finity!

Thanks for the read! Let me know what you thought by replying back to this email.

— Alex