Over the past 72 hours, a single event switched off payment cards across at least two independent wallet ecosystems. Ready, the self-custody wallet formerly known as Argent, terminated its card program after Kulipa, its issuing partner, abruptly wound down. Founder Itamar Lesuisse says no advance notice was given; users learned at the same moment as the team. Solflare, the Solana wallet, was affected as well. This is not a smart-contract exploit. It is an off-chain dependency failure wearing the clothes of a Web3 product. Code does not lie, only the architecture of intent.
Let me be precise about what happened. Ready is a smart-contract-based self-custody wallet with roots in the ZKsync and Starknet ecosystems. The wallet's value proposition is user sovereignty: funds remain on-chain, keys stay in the user's hands, and the wallet acts as a gateway to DeFi. The card program was meant to connect that sovereign asset layer to the fiat world. A user would deposit stablecoins, receive a card, and spend through conventional card rails. Kulipa was the licensed issuer responsible for KYC, compliance, BIN sponsorship, and settlement. It was the pipe connecting crypto-native wallets to the traditional payment network.
A crypto card issuer sits between the wallet and the traditional financial system. It is not merely an API wrapper; it is the holder of the BIN, the party accepted by the card network, and the entity that faces the bank sponsor. For Ready, Kulipa handled the flow of compliance data, transaction authorization, and settlement. When an issuer winds down, the normal lifecycle includes a transition period. In this case, there was no transition. The result was not a partial outage; it was a full termination. Multiple products built on Kulipa's rails went dark in one day. This is what a hidden single point of failure looks like in a supposedly decentralized industry.
Self-custody vs. service continuity
Self-custody protects the asset layer but not the payment interface. The user's on-chain balance was never controlled by Kulipa. In that narrow sense, the self-custody thesis passed. But the card was not an asset; it was a service. Services carry counterparty risk. When Kulipa stopped, every card user lost the ability to spend. The function disappeared. A user staring at a safe wallet balance and a frozen card is not a user who has been protected; they are a user who has been reminded that the crypto economy still depends on centralized fiat rails.
The shared issuer hidden in the dependency graph
Solflare and Ready are separate teams with different chains, users, and security models. They shared one issuer. This is the same single-point-of-failure pattern I saw in oracle design during the 2020 DeFi summer. In my audits of lending protocols, I warned that a single price feed could produce cascading liquidations if it became stale or manipulable. Kulipa is the price oracle of the card world: a shared critical component whose health few teams could monitor. The dependency graph matters more than any individual smart contract. If you review a wallet project and its security model does not include the card issuer, you are reviewing half of the product. I have spent the past decade assuming that third-party dependencies will eventually break. This incident is the latest proof.
Replacing Kulipa is not a configuration change
Replacing Kulipa requires more than swapping an API endpoint. The new issuer must receive a new BIN, complete banking-network certification, implement its own KYC/AML stack, and sign sponsorship agreements with financial institutions. Users must re-submit identity documents, wait for approval, and receive a new physical card. Existing card numbers die. Subscription merchants see declined payments. Direct debits collapse. Based on my work integrating financial infrastructure, this process takes months. The announcement 'we are looking for a new issuer' is not a recovery plan; it is the beginning of a long operational migration.
Accounting opacity
Ready stated that user funds were unaffected. I interpret that as the smart contract assets under self-custody were not touched. It does not tell us about stablecoins pre-loaded into the card, settlement batches held by Kulipa, or refunds stuck in a payment processor. Without an audited reconciliation from the issuer, every dollar inside the payment pipeline is unknown. Truth is found in the gas, not the press release. One caveat: the only source for 'user funds unaffected' is a founder statement. It is a reasonable presumption for self-custodied funds, but it is not an audited proof. Users should check their own balances on the block explorer. That is the only on-chain fact available. Everything else is narrative.
The Contrarian Angle
The contrarian angle is not that Kulipa was a centralized point of failure. Every card program is centralized. The true blind spot is the industry's tendency to equate 'funds are safe' with 'the product survived.' It did not. For a user who relied on the card for daily life, the ability to spend is an asset even if it never appears on a balance sheet. An unspendable balance is a form of locked liquidity.
Another blind spot is the failure of the 'don't trust, verify' doctrine outside the chain. In DeFi, you can verify a smart contract, inspect the bytecode, and simulate transactions. There is no equivalent for a BIN license or a bank sponsor. The issuer's relationship with its banking partner is a private contract. Its regulatory filings may not be public. Its risk models are invisible. The moment a product depends on off-chain infrastructure, the verification domain shrinks to a press release. That is why the industry needs something less elegant: redundant fiat exits, issuer health metrics, and a fallback that does not require a new plastic card. Hedging is not fear; it is mathematical discipline. If the logic is not sound, the interface does not save you.
Market Consequences
This event will not move Bitcoin or Ethereum. It is too small to be a macro factor. But it is a meaningful data point for anyone tracking the institutionalization of crypto payments. The narrative that self-custody wallets are ready for mainstream consumers was always conditional on their ability to connect to the traditional payment network. That connection is the weakest link. A person who used an Argent card for coffee and rent now knows that the crypto card is not a bank card; it is a crypto card with a bank-shaped accessory. The market will not reprice a token, but it may reprice the entire 'wallet-to-payment' category. Institutions building on this stack will begin asking for issuer redundancy in due diligence. That is the actual information gain from this incident.
The competitive effect is subtle. Exchange-issued cards from Binance or Crypto.com wrap custodial balances; their issuers are usually subsidiaries of the exchange itself. They have a different risk profile because they control both the assets and the card. Self-custody wallets cannot own a bank or a card network, so they outsource to specialized issuers. This incident gives exchange cards a short-term advantage in reliability, but it also validates the self-custody security model: users did not lose funds. The question is whether users value control or continuity. The answer is likely both, and that demands a new type of infrastructure. It demands a multi-issuer card abstraction that can fail over from one licensed provider to another without forcing users to re-KYC or change spending habits.
Regulation will shape the fix. If Kulipa's shutdown was caused by a bank sponsor withdrawing, no purely crypto-native solution can prevent the next one. The issuer's license is tied to banking relationships that are outside the wallet's control. That is why the solution is not to buy a bank but to create a standard for issuer transparency: audited solvency reports, on-chain proof of settlement balances, and a public registry of card program status. Without such a standard, every card program will be one phone call away from termination.
What Needs To Change
Wallet teams should treat every external dependency as a potential point of collapse. The immediate priority is a manual audit of card issuer, payment processor, banking partner, and compliance vendor. Each dependency should have two independent alternatives, and those alternatives should be tested on a simulated card transaction. Next, wallet teams should expose an issuer-health dashboard to users. If a card program is down, the user should see which layer failed and what the recovery path is. Legal agreements should also include mandatory notice periods and contingency plans. None of this is exciting. That is the point. Simplicity is the final form of security.
For users, the immediate checklist is simple: export your transaction history, check whether any balance is held inside the card rather than on-chain, cancel automatic payments linked to the card, and treat 'funds are safe' as a claim to verify, not a fact to accept. Do not assume that a new card will arrive in days. Plan for a three-month off-ramp gap. This advice is deliberately unglamorous. The most dangerous error in a vendor collapse is the assumption that the system will fix itself.
Takeaway
The next 90 days will determine whether this is an outlier or a systemic signal. If a second card issuer winds down without notice, the market will correctly treat self-custody card programs as fragile infrastructure. The response should be architectural, not rhetorical. Wallets need a multi-issuer abstraction layer that routes around a dead partner, in the same way decentralized networks route around a failed node. Simplicity is the final form of security. Until a wallet can show two independent fiat exits, the self-custody story remains half-built. History is a dataset we have already optimized, and it keeps returning the same result.