China's Credit Card Slump 2026 and the Virtual Card Boom

The numbers first: the 700 million card milestone broke

Card issuance in China has now fallen for three straight years. Per the central bank's payment system data, the national credit card count dropped below 700 million at the end of 2025, the lowest level in seven years. 21st Century Business Herald counted in April 2026 that banks including Bank of Communications, Minsheng, China Guangfa and Agricultural Bank had pulled more than 42 credit card products in the first four months of the year alone. CCTV Finance put the figure above 45. Almost all of them were co-branded or themed cards.

The channel side is contracting even harder. Sixty-six credit card branch centers closed in 2025, fifty-eight of them at Bank of Communications alone. At least nine more shut in early 2026, with Guangzhou Bank approved to close seven less than a month into the year. Several banks have also folded their standalone credit card apps into their main mobile banking apps, so the entry points are narrowing too.

Asset quality tells the same story. ICBC's credit card NPL ratio climbed to 4.61%, up 111 basis points from 3.5% a year earlier. Among the six state-owned banks, only Postal Savings Bank managed to cut both its ratio and its balance. Volume down, branches closed, risk up. That is the credit card business in 2026.

Why banks are cutting cards on purpose

This is not banks losing interest. It is banks that can no longer afford to issue freely, for four stacked reasons.

Reason one is a regulatory red line. The 2022 joint notice from the banking regulator and the central bank capped long-term dormant card rates at 20% and banned using issuance volume as a primary KPI. For a decade banks had pushed card counts as the core metric, branch staff carried quotas, underwriting loosened layer by layer, and the system accumulated hundreds of millions of sleeping cards. Once the rule bit, those cards got purged from the statistics and issuance numbers collapsed. A UnionPay data consultant described it plainly: the industry is squeezing water out of the numbers, not collapsing.

Reason two is that the co-branding model stopped working. A co-branded card borrows an IP to drive signups, and IP heat is cyclical while contracts expire. When the heat fades, transaction frequency drops with it, and keeping a low-activity card alive burns marketing budget forever. Minsheng killed eleven products in one announcement. Guangfa killed five. Same math every time.

Reason three is delinquency. Card NPLs are one slice of the retail risk now surfacing across banks, so lenders are pushing out high-risk segments and tightening approval. In plain terms: getting a new card is harder than it used to be, especially for freelancers and anyone with unstable income.

Reason four is channel consolidation. Standalone card apps cost money to run, folding them into mobile banking is the trend, and the branch centers follow. Stack the four together and this looks less like a dip and more like a multi-year slimming program.

Who actually gets hurt

First, kill the rumor. The viral claim that credit cards are finished in China is clickbait. Existing cards keep working, and holders of discontinued products get switched to a standard card of the same tier at reissue. The perks die, the card does not. Multiple banks confirmed this to Xinhua.

The real impact lands on three groups. People who wanted a card and now face tighter underwriting. People who liked co-branded perks and watch those expire with fewer replacements on the shelf. And people with cross-border payment needs, where the effect is hardest to see but runs deepest. Domestic cards were already mostly useless for subscriptions like ChatGPT and Claude, because issuer-country risk filtering rejects Chinese BINs at the gateway. Now even the path of getting a card first and worrying later is narrowing.

One background number deserves more attention than it gets. China has over 300 million AI users, and effectively none of them can pay for a US AI subscription directly with a domestic bank card. The credit card contraction builds another wall beside that gap, and the wall is still growing.

Why virtual card demand runs the other way

The demand logic is simple. Cards are getting harder to issue, but cross-border spending has not slowed down. AI subscriptions, ad accounts, overseas e-commerce, software renewals, all of it needs a card that clears a foreign-currency rail. Shrinking supply meets rigid demand, and the gap gets filled by instruments whose BINs sit outside China.

Prepaid virtual cards sit exactly in that slot. No credit check, no credit line consumed, low entry barrier. Complete KYC, fund the wallet with USDT or whatever the platform accepts, and you have a Visa or Mastercard number in minutes. For someone who cannot get approved for a credit card, or refuses to apply for one just to pay a 20-dollar-a-month subscription, this is the lowest-friction path on the market. Platforms like PokePay and FotonCard have spent the past two years growing on precisely this structural gap.

One thing needs to be said clearly though. A virtual card is a prepaid instrument, not a credit instrument. Its limit equals whatever you loaded onto it, nothing more. It does not replace a bank's credit line. What it replaces is that one specific card you could never obtain for cross-border scenarios. Domestic spending, installments, emergency credit, the bank is still the right answer there.

Worth adding: the platforms themselves carry risk. This industry has buried plenty of names over the past two years. Before funding any wallet, check the license, the custody arrangement and the operating history one by one. Our license-tier and fund-custody verification guide covers the full method.

Action lists for three reader types

If you already hold a credit card: keep it. Check whether international card-not-present payment is enabled, because many banks ship it disabled and it takes two taps in the app to switch on. If your dual-currency card still clears foreign merchants, think hard before canceling it. New cards will be harder to get and the limits more conservative.

If you cannot get a card: pick virtual cards by use case. For AI subscriptions only, the selection logic in our AI subscription payment guide applies, and US-BIN Mastercard clearly outperforms Visa in acceptance. For a general-purpose US card, the MXK8 walkthrough gets you a Visa in about five minutes. Read the decline troubleshooting guide before you start, it saves a lot of trial and error at checkout.

If you run ad accounts or a business: corporate virtual cards are the standard answer. Google and Meta ad platforms are acutely BIN-sensitive, and one contaminated shared BIN segment can take down every card issued on it. Card selection for ad spend follows completely different logic, which we covered separately in the ads payment guide.

Three questions we keep getting

Will China stop issuing credit cards entirely? No. The cuts concentrate in co-branded and themed products. Standard card lines continue, and China CITIC actually added roughly six million cards in 2025. What is being culled is inefficient product, not the category.

Will regulators ban virtual cards? A blanket ban does not look likely near term, but the compliance direction is unmistakable. The US GENIUS Act already treats stablecoin issuers like financial institutions, KYC costs are flowing into top-up fees, and the no-KYC niche keeps shrinking. Preferring licensed platforms is essentially buying regulatory cover.

Do I need to act on a discontinued card? Not really. Expired perks do not affect the card's basic function. The one thing worth checking is whether any recurring subscription bills to that card, because a reissue with a new number can quietly break the billing chain. Our piece on the VAU account updater mechanics breaks down exactly how that happens.

A sober judgment

Credit cards are not dying. The crude issuance model is. What fades is the plastic, while credit consumption itself moves deeper into digital channels, and I buy that framing. Shanghai Pudong Development Bank grew its card loan book more than 5% against the trend by betting on EV installment lending. CITIC grew issuance while the industry shrank. Banks are not retreating, they are trading low-value cards and branches for high-value customers.

The virtual card boom is a mirror of that structural shift. Domestic credit consumption digitizes and gets refined, while foreign-currency spending gets pushed onto prepaid instruments with overseas BINs. As long as overseas subscriptions and cross-border business exist, the gap exists. The variable actually worth watching is regulation: KYC cost pass-through, stablecoin rulebooks, and slower BIN approvals upstream are all quietly raising the price of using a virtual card. Getting your accounts and compliance posture sorted while the window is open costs far less than fixing it after the wind changes.