The End of the CEX Supercycle: What's Happening with Crypto Exchanges in 2026
Depositing money on crypto exchanges has long become routine — both within the affiliate industry and beyond it. In the past, exchanges would occasionally shut down or fall victim to hacks. But the current situation is different: three well-known brands closed within a single month, and one of them managed to collect $40 million from users before pulling the plug.
In this article, we break down what's happening in the tier-2 CEX market and whether the problem is systemic.
The structure of centralized crypto exchanges
The centralized exchange market can be broadly divided into two tiers. The first — Binance, OKX, Kraken, Bybit, Coinbase: billion-dollar volumes, regulatory reporting, public proof of reserves.
The second tier — dozens of mid-sized exchanges competing for users with smaller budgets, often without transparent reporting and with far less resilience to withstand a liquidity crunch.
As history shows, the biggest problems tend to emerge among tier-2 exchanges — though there have been scam cases at first-tier projects as well. The loudest wave hit in 2022: first, the TerraUSD stablecoin collapsed, taking the associated Luna token down with it.
Hedge fund Three Arrows Capital failed to cover its positions and went bankrupt, and through a chain of uncollateralized loans dragged several major players down with it. Voyager Digital froze trading on July 1, 2022, and filed for bankruptcy within days. A week earlier, lender Celsius Network did the same. In November of that year, FTX collapsed — at the time the world's second-largest exchange by volume, with a hole of billions of dollars in client funds.
According to estimates from the Federal Reserve Bank of Chicago, FTX, Celsius, Voyager, BlockFi, and Genesis collectively wiped out $46.5 billion and affected 4.3 million users in just five months. Chicago Fed researchers noted a common pattern: all five companies offered instant withdrawals while holding assets in illiquid, high-risk instruments in pursuit of unrealistic yields.
Since then, the list of dead, frozen, or liquidated exchanges has only grown — Cryptopia, QuadrigaCX, Mt. Gox, and dozens of lesser-known names. In September 2024, German authorities seized the infrastructure of 47 no-KYC exchanges as part of a money laundering crackdown.
The pattern was the same in every case — a mid-sized exchange takes on obligations it cannot meet during a mass withdrawal event, and stays afloat on the flow of new deposits. When that flow stops, the point of no return is reached.
The mass summer shutdown of 2026
July 2026 turned out to be a busy month for centralized crypto exchange closures. Where such events used to happen once every few years, this time three well-known mid-tier exchanges announced they were winding down operations within a single month.
The first was AscendEX — on July 1, the company announced it was ceasing operations and beginning the wind-down process. The exchange offered no clear guarantees regarding full reimbursement of funds, and the withdrawal process remained opaque.
AscendEX did not disclose the volume of funds left frozen. Meanwhile, on-chain analysts had been tracking a rapid deterioration in the exchange's financial health even before the closure: on June 20 alone, over $240 million was withdrawn from its wallets, and by early July public reserves had shrunk to $13.5 million — of which more than $12 million consisted of the illiquid ASD and UNITE tokens.
The second was BitMEX. On July 23, the exchange — which had been operating in the market since 2015 — announced it was shutting down. Unlike AscendEX, the company published a wind-down schedule and a phased asset withdrawal plan.
The exchange reported no liquidity issues or fund shortfalls, and at the time of the announcement claimed to hold over $739 million in client assets. Users were given nearly two months to withdraw their funds, with the wind-down proceeding according to a pre-published schedule.
The most high-profile closure was BitMart, which on July 26 announced an "orderly wind-down." This case most closely resembles a classic exchange collapse driven by a liquidity deficit.
According to the fired CEO and employee testimonies, there was no "planned wind-down" at all. In April, a hole was discovered in the balance sheet, after which management began restricting large withdrawals while funding smaller payouts from incoming deposits. To attract more capital, the exchange launched an unlimited deposit product — initially offering 12% APY, later raised to 18.88%. In this way, a platform that was effectively already insolvent managed to raise approximately $40 million from users.
On-chain data was sending alarming signals at the time. BitMart's wallets held approximately $2 million, while outflows over just one month amounted to around $10 million. Despite this, the exchange continued to report $1.3 billion in daily spot volume in its public statistics. The CEO was let go two days before the official announcement and, by his own account, learned about the closure from the same notice as regular users.
The contrast between public statements and the actual state of affairs is particularly telling. On July 17 — nine days before the shutdown announcement — BitMart published its H1 report, claiming a 256% increase in assets under management. Following the news of the closure, the BMX token collapsed 58–81.5% within a week.
This is already the second major incident in BitMart's history. In 2021, the exchange was hacked for $150–200 million, and at the time the founder claimed to have covered the losses out of his own pocket. In 2026, the gap was effectively funded by users who had deposited money chasing elevated yields: as part of the wind-down procedure, they were offered compensation of 100 USDT per coin with a 10% withdrawal fee.
In total, AscendEX, BitMEX, and BitMart all exited the market within a single month. According to data from on-chain analyst 0xviet, as of the end of July 2026, at least 63 crypto projects had announced closures — ranging from L1 and L2 networks to DeFi protocols and centralized exchanges.
How to approach crypto exchanges?
If the crypto exchange collapses of 2022 were driven by systemic shock following the Terra and FTX implosions, the summer of 2026 tells a different story. The problem is increasingly rooted in the business models of the platforms themselves.
Most mid-tier CEXs spent years competing not on technology or liquidity, but purely through aggressive marketing: high deposit rates, bonuses, and referral programs. As long as new users kept flowing in, the model held up.
But following the launch of spot Bitcoin ETFs, tightening regulation, and the concentration of liquidity around a handful of major exchanges, the influx of new clients slowed significantly. For platforms that depended on constant deposit growth, this became a serious problem.
That is precisely why the BitMart story reads as a textbook symptom of the end of the tier-2 CEX supercycle. When an exchange starts offering something like 19% APY with no deposit cap, it signals a desperate need to attract fresh capital. If on-chain metrics are deteriorating at the same time, reserves are shrinking, and large withdrawals are starting to face delays, the risks to users escalate sharply.
Reported trading volumes are no longer a sufficient indicator of a platform's financial health. BitMart continued to show approximately $1.3 billion in daily spot volume even as its publicly identifiable on-chain reserves were estimated at just around $2 million.
This is why independent on-chain data, Proof of Reserves, and reserve management transparency are becoming increasingly critical — they make it possible to assess whether an exchange is genuinely capable of meeting its obligations to clients.
DEXs are beginning to outpace CEXs in volume
Conclusions
Whether July 2026 marked the end of the supercycle for CEXs remains to be seen. But it is already clear that the era in which second-tier exchanges could buy trust through high yields and marketing is drawing to a close. The market is becoming less forgiving of weak business models, and the cost of failure for users remains unjustifiably high.
