When a crypto payment provider shuts down overnight, custody decides who loses money
Key takeaways
- On July 29, 2026, the Paris-based stablecoin card issuer Kulipa stopped operating over solvency problems. Cards declined the same day, with no wind-down window, across about 20 client platforms.
- The self-custodial card programs built on it lost no balances. Funds left the user's own wallet only at the moment of purchase, so nothing sat with the provider.
- Funding is not a safety signal. Kulipa had raised $9.2M in total, including a $6.2M seed announced on April 1, 2026, and closed about four months later.
- A merchant has more at stake than a card user. Unpaid sales, a rolling reserve and queued payouts can all sit with the provider at once. In the worked example below, a $60,000-a-month shop has $32,000 exposed.
- With a non-custodial processor, the customer's payment lands in the merchant's own wallet. There is no provider balance to freeze.
What happened when Kulipa stopped on July 29
On July 29, 2026, Kulipa ceased operations. It did not announce a wind-down for September. It stopped. Cards that worked at breakfast were declining by the afternoon.
Kulipa had no end users. It ran the backend of other companies' card products: processing, fraud management and settlement. That let a wallet ship a card without becoming a card company.
The Block reported on April 1, 2026 that Kulipa had raised a $6.2 million seed, co-led by Flourish Ventures and 1kx. Total funding reached $9.2 million. The company claimed more than 120,000 cards issued since its February 2025 launch, 20 signed customers and 70% month-over-month volume growth. It shared no absolute figures.
Four months later it was gone. PANews reported that Kulipa told clients only that it had a "solvency issue", that it reportedly looked for a buyer without success, and that its site went offline. An analyst quoted there blamed thin margins and a model that depended on the next funding round.
The market was not shrinking. The same report puts crypto card spending at about $1.5 billion a month by the end of 2025. It cites The Kobeissi Letter for $7.8 billion in May 2026, up 230% year over year. A growing segment does not save every provider in it.
Why the self-custodial cards lost no balances
Solflare co-founder Vidor Gencel posted the same day that the issuing partner was winding down over solvency issues. He apologised for how users found out. Then he made the point that mattered: funds are safe and there is nothing you need to do.
The card was self-custodial end to end. There was no deposit and no top-up. Nothing was parked with the card provider. The card spent from the user's own wallet at the instant of purchase.
Ready (formerly Argent) was built the same way. Both wallets could say within hours that user assets were untouched. PANews reports that nSave, a custodial fintech client, negotiated a short delay, told its users and also reported no lost funds. Flutterwave had made no public statement when PANews published.
So about 120,000 cards stopped working, and the coverage cited here reports no lost balances. In the self-custodial programs, one design choice made long before the shutdown turned a loss into an outage.
A shutdown on a $60,000-a-month shop's balance sheet
A consumer whose card dies has a bad afternoon and reaches for another card. A business is in a different position. Its processor holds revenue for a while, then sends it on.
Take a shop with $2,000 of crypto sales a day, or $60,000 over a 30-day month. Assume a custodial provider pays out weekly and keeps a 10% rolling reserve for 90 days. These are illustrative terms, not any named provider's. Here is what sits with the provider on an ordinary day:
| Money at the provider | Calculation | Amount |
|---|---|---|
| Sales not yet paid out (worst day, just before payout) | 7 days x $2,000 | $14,000 |
| Rolling reserve held | 10% x $2,000 x 90 days | $18,000 |
| Total exposed | $14,000 + $18,000 | $32,000, or 16 days of takings |
| Held by Payzum | Settlement goes to the merchant's wallet | $0 |
Payzum charges 0.5% per processed sale, deducted from the amount received, plus on-chain network gas. On $60,000 that is $300 a month.
The $32,000 would not vanish, but it would stop being yours to spend. How fast it comes back is set by the estate, not by you. Celsius, a crypto lender and not a processor, shows how slow that can be. It paused withdrawals in June 2022 and filed in July 2022. On January 4, 2023 a judge ruled that the crypto in its Earn accounts, about 600,000 accounts worth $4.2 billion, belonged to the estate, CoinDesk reported. By late August 2024, Cointelegraph reported, it had paid about 251,000 of more than 375,000 eligible creditors. Terms of use decided who owned the funds, and the wait ran for years.
The cash is only part of it. You also lose checkout for a few days, staff improvise at the counter, support tickets pile up, and you migrate to a replacement you had no time to evaluate. If you also pay out affiliates or contractors, that side freezes too.
None of this takes fraud or bad faith. A company holding your money only has to run out of its own.
Why "regulated and well funded" does not settle it
After a shutdown the instinct is to find a bigger, better-capitalised provider. That is reasonable and not enough, for three reasons.
Funding looks backward. Kulipa closed a seed round and stopped operating in the same year. A full bank account at announcement time says something about investors months ago. It says nothing about next quarter's runway, margins or compliance costs.
Risk also hides in layers you never signed with. Your provider may depend on an issuer, which depends on a partner bank, which answers to its own regulator. Kulipa's clients did not choose its dependencies. They inherited them.
Contracts decide who is eventually owed what. They do not put money in your account on the morning the platform goes dark. Recovery through an estate is measured in quarters, as the Celsius timeline above shows.
So the lasting question is not how strong the company is. It is what the company's failure does to you. Architecture fixes that answer, and you can check it before you sign. If you are comparing providers on this point, the Payzum vs Coinbase Commerce page sets out the custody models side by side.
What non-custodial means for a merchant
Payzum is a non-custodial, crypto-only payment processor. The customer's payment goes directly to a wallet the merchant controls. Payzum never holds, pools or controls the money. There is no merchant balance in a Payzum account and no payout cycle, so no pile of your revenue is exposed to Payzum's fortunes.
Run the Kulipa test on it. If Payzum stopped tomorrow, you would lose the tooling: checkout, dashboard, QR generation, webhooks, reporting. That is a real problem and a migration you would rather skip. You would not lose settled revenue, because none of it passed through us. Tuesday's sales are in your wallet on Tuesday.
Things can still go wrong, but your money going missing with the vendor is not one of them.
On the acceptance side, it looks like this:
- In person: a POS that generates a new QR per sale. Any phone becomes a terminal, cashiers log in with PINs, and you get analytics per cashier and per terminal. There is no acquirer, no card-network fee and no chargebacks.
- Online: hosted checkout (redirect, modal or inline), no-code payment links and buttons, invoices with expiry and overpayment detection, recurring subscriptions and donation flows. They work with existing e-commerce plugins, snippets and webhooks.
- Volatility: accept what the customer holds, with optional auto-convert to USDC or USDT. See USDT vs USDC for payments.
- Payouts: mass payouts by CSV (BTC, LTC, DOGE) and EVM stablecoin payouts on Polygon, Arbitrum, Optimism, Base, BNB Chain and Avalanche, so the outbound side does not wait in someone else's queue.
- Networks: Bitcoin, Ethereum, Solana, Polygon, Base, Arbitrum, Optimism, BNB Chain and Avalanche.
Moving settlement to your wallet in four steps
- Point settlement at a wallet you already control. You give a destination address, or one per network, and complete KYC in the product. You are not opening an account that holds your money.
- Turn on the surfaces you sell through: POS for the counter, payment links for WhatsApp or Instagram, hosted checkout for the store, invoices and subscriptions for recurring clients. Pick networks and coins, and whether to auto-convert.
- Take the payment. At the till, a fresh QR for that exact amount. Online, a hosted checkout or a link. The funds arrive in your wallet, not in a pending balance.
- Reconcile. Signed webhooks push events into your systems, the REST API covers anything custom, and audit logs, 2FA and per-cashier analytics help you close the day.
No step asks you to wait for someone to release your funds.
Who a provider failure hurts most, and what changes
- The counter business with thin margins. A café, barbershop or auto shop cannot absorb a week without takings. With settlement landing per sale in the merchant's wallet, an outage means finding another way to make a payment request. It does not mean chasing a receiver for last week's revenue.
- The cross-border service business. An agency or studio invoicing clients abroad knows money can sit in intermediaries. An invoice paid in USDC arrives in the wallet directly, with overpayment detection and expiry handled.
- The operator that pays people out. Marketplaces, affiliate programs and creator platforms owe money in both directions. Mass payouts from your own wallet mean a provider's problems do not become your contributors' problems.
- The API provider selling to machines. With x402, agents pay USDC on Base per call. Payzum is the middleware in front of your existing endpoint: you configure the endpoint, your API key and a price. An external facilitator, currently Coinbase's, settles the payment before the paid call is proxied through. Revenue arrives per call, with no balance building up anywhere.
Custodial provider vs non-custodial processor when things go wrong
| Scenario | Custodial provider holds your funds | Payzum (non-custodial) |
|---|---|---|
| Provider becomes insolvent | Settled balances sit in the estate; you are a creditor | No provider-held balance exists; funds are already in your wallet |
| Provider suspends your account | Access to your balance and payouts can be blocked | Nothing to block; past settlements are in your wallet |
| Provider has an outage | Checkout and access to funds can both stop | Checkout tooling stops; already-settled funds are unaffected |
| Upstream partner (bank or issuer) fails | Inherited dependency you never signed with | Settlement is on-chain, to your address |
| Time to get your money after a sale | Payout cycle, plus any rolling reserve | Settlement is the payment |
| Disputes and reversals | Chargebacks reversible for months | On-chain finality: no chargebacks |
Common objections
"If Payzum can also disappear, what's the difference?"
Any company can disappear. That is the honest baseline, and it is why the question is what a failure costs you. If a custodial provider fails, you lose the tooling and the money it held. If Payzum failed, you would lose the tooling and migrate. The revenue is already yours, in a wallet whose keys we never had.
"Doesn't holding my own wallet just move the risk onto me?"
It moves the risk to a place you can control with process: a hardware wallet or multi-sig for treasury, a separate operating wallet, documented key custody and staff access rules. That is security discipline with known best practice. You cannot mitigate a counterparty's solvency. You can only find out about it, usually late.
"We're small. Isn't this a problem for big merchants?"
It is the other way round. A large merchant has lawyers, negotiated terms and several processors. A five-person business with one provider and two weeks of settled revenue in flight is the one for whom a shutdown is existential. The $60,000 shop above carries 16 days of takings at risk.
"Do I have to give up cards to do this?"
No. Payzum is crypto-only and does not replace your card acceptance. It adds a rail where the money reaches you directly and cannot be reversed. Many merchants run both.
Frequently asked questions
What happens to my money if my crypto payment provider shuts down?
It depends on custody. If the provider held your settled funds, those balances become part of an insolvency process and you recover as a creditor, which typically takes months or longer. If the provider was non-custodial and payments settled directly to a wallet you control, you lose access to the tooling but not to the funds, because the provider never held them.
What exactly happened with Kulipa?
Kulipa was a Paris-based company providing white-label stablecoin card infrastructure to wallets and fintechs. On July 29, 2026 it ceased operations over solvency issues, and cards stopped working the same day across roughly 20 client platforms, including Solflare, Ready (formerly Argent), Flutterwave and nSave. It had issued more than 120,000 cards since February 2025 and had raised $9.2 million in total, including a $6.2 million seed announced in early April 2026.
Did Kulipa's users lose their stablecoins?
Not in the self-custodial programs, and the coverage cited here reports no lost balances elsewhere. Those card programs were self-custodial end to end: there was no deposit or top-up, and no user balance was held by the provider. Funds left the user's own wallet only at the moment of purchase, so when the provider stopped operating the cards became unusable but the balances stayed exactly where they were.
How is Payzum different from a custodial crypto payment processor?
Payzum never holds, pools or controls merchant funds. The customer's payment settles directly to a wallet the merchant controls, so there is no Payzum balance, no payout cycle and no rolling reserve. Payzum is crypto-only and does not settle to fiat bank accounts; optional auto-convert to USDC or USDT is available for volatility protection.
Is self-custody safe enough for a business?
It requires discipline rather than trust: a hardware wallet or multi-sig for treasury funds, a separate operating wallet, documented key procedures and clear staff access rules. The trade-off is that your risk becomes operational and controllable instead of dependent on a third party's balance sheet, which you cannot see or influence.
Can I still accept payments in person if I go non-custodial?
Yes. Payzum's POS generates a new QR per sale, turns any phone into a terminal, supports physical terminals and PIN logins for cashiers, and provides analytics by cashier and terminal. There is no acquirer in the flow, no card-network fees on acceptance and no chargebacks.
Check who holds your money before you need to know
Kulipa's clients found out on a Wednesday afternoon, at a checkout, with no email. Those that had chosen self-custody earlier only had to explain an outage. Book 20 minutes with the Payzum team and we'll go through your actual payment flow: where funds sit, for how long and with whom. Then we'll design the non-custodial version of it for your business.
Prefer email or a different time? Pick a slot here · [email protected]
This article is analysis of public reporting and public statements, not legal, financial or tax advice. Details of the Kulipa shutdown are as reported by third parties as of August 3, 2026 and may be updated by later disclosures. Confirm the rules and obligations that apply to your business with a qualified professional in your jurisdiction.