Gauntlet and the trouble with too much money

On October 30, 2025, two lending vaults curated by Gauntlet received a $775 million deposit. Ordinarily, this is the part of a startup story where somebody orders champagne. Instead, the annualized yields fell by more than 70%. Money had arrived faster than anyone wanted to borrow it. The celebration had encountered arithmetic.

That awkward afternoon is a useful introduction to the company. Crypto likes to describe finance as a software problem. Gauntlet occupies the place where software meets the less obliging realities of borrowers, collateral, and people who disagree. Its product is judgment, expressed through models and turned into allocations.

A deposit is not a borrower

In its February 2026 account of the event, Gauntlet says yields recovered within ten days. The response combined new collateral markets, revised allocations, and work with Morpho to attract borrowing. Borrowed capital rose from $89 million to $716 million. This was a company-reported case study, but its mechanism is refreshingly ordinary: lenders earn interest when borrowers pay it.

The practical lesson travels beyond crypto. Before celebrating new supply, work out who will use it. A marketplace needs buyers; a credit product needs borrowers. An allocation engine can distribute money elegantly and still distribute it into markets with insufficient demand.

The simulation becomes the decision

Founded in 2018, Gauntlet brought quantitative trading methods to crypto economics. Co-founders Tarun Chitra and Rei Chiang came with experience in quantitative research and high-frequency trading. John Morrow, also identified publicly as a co-founder, helped build its operating business. The early proposition was to test financial systems before the market tested them more expensively.

Its research asks questions a code audit alone cannot settle. How much can users borrow against a token? What happens if its price falls while trading liquidity disappears? Can liquidators sell collateral quickly enough to cover debt? A contract can execute exactly as written while its economics produce an unpleasant answer.

Gauntlet Applied Research still sells that expertise. Its work spans risk parameters, incentives, liquidity programs, and mechanism design. Compound and Coinbase’s Verified Pools appear among its public engagements. An operator can use the service to examine whether a rewards budget buys useful liquidity, or merely rents it until the rewards stop.

The committee had other ideas

For years, Gauntlet advised Aave on market risk. A December 2023 renewal approved a $1.6 million annual fee, split between GHO and aUSDT. Thirty percent was assigned to an insolvency-refund vault. This was a specified engagement with specified payment terms, rather than an invitation to pay whatever a dashboard suggested.

Then, in February 2024, Morrow announced Gauntlet’s departure. He cited difficulty navigating stakeholders’ “inconsistent guidelines and unwritten objectives.” The statement describes a relationship that had become difficult to operate. It also reveals a constraint on the advisory model: the people calculating a risk limit may lack authority to implement it.

“inconsistent guidelines and unwritten objectives”John Morrow / Aave departure announcement, February 2024

The vault business offered a different arrangement. On Morpho, a curator chooses the markets a vault lends into and manages exposure within its permissions. Depositors choose the strategy. Gauntlet could apply its research through allocations, with greater operational responsibility attached. The appeal of that arrangement is clear; treating the Aave dispute as its sole cause would oversimplify the story.

Three appetites, one lending machine

Gauntlet’s lending lineup makes the tradeoffs explicit. Prime targets relatively conservative exposure. Balanced mixes larger and smaller collateral markets. Frontier reaches into potentially more volatile, less liquid opportunities. A higher yield target comes with a reason, and the reason deserves at least as much attention as the percentage.

USD Alpha broadens the approach into a stablecoin strategy built on Aera infrastructure. It rebalances across opportunities and reinvests rewards. Custom vaults let partners build their own offerings; levered real-world-asset strategies add borrowing against tokenized assets. These are different financial exposures, even when the interface makes them look equally tidy.

The business earns performance fees from interest generated by selected vaults, alongside paid research engagements. Morpho’s account notes that some vaults have zero performance fees. There is no universal sticker price: prospective users need the particular vault’s fee terms, available liquidity, and withdrawal mechanics. USD Alpha’s documentation says withdrawal requests may occasionally take up to twelve hours.

Gauntlet sits between the protocol supplying the machinery and the customer supplying the money. Other Morpho curators, including Steakhouse Financial, offer alternative allocation judgments. Gauntlet’s particular proposition combines simulation research, several risk mandates, and ongoing management. The comparison worth making is between actual portfolios and operating processes. A curator’s name alone tells you little about what sits beneath a deposit.

The risk travels through the button

Distribution now runs through familiar interfaces. Gauntlet-curated Morpho vaults power Uniswap Earn. Raenest added stablecoin yield in September 2026; DPT integrated USD Alpha into its payments app. Institutions can obtain allocation expertise, while individual users can deposit through supported apps. Neither group needs to become a full-time researcher to use the product.

They still inherit its dependencies. After Resolv’s March 2026 USR exploit, Gauntlet reported removing related markets, including exposures in Core and Frontier vaults. Its update discussed roughly $7.6 million and $4.3 million in affected liquidity. Those figures were not a final loss accounting. The episode shows how an underlying token’s failure can reach a curated lending strategy.

For a prospective integrator, the useful checklist is short: inspect collateral, concentration, fees, and withdrawal conditions; understand who can change exposure; examine incident handling. The approach needs dependable data, functioning price feeds, sellable collateral, and borrowers. When those conditions deteriorate, a risk model cannot manufacture an exit.

SBI Holdings backed a $125 million Series C announced in July 2026. That gives Gauntlet money to extend its institutional business. The more interesting test remains the old one: can it put capital to work without mistaking an attractive number for a sound loan? Behind the effortless button, somebody still has to ask the awkward questions.