The number MEXC wants carried forward is 297. That is the Bitcoin reserve ratio in its September 2026 Proof of Reserves report — audited by Hacken, snapshot-dated September 10, and released five days later from Mutsamudu, Comoros. A 297% ratio reads like a fortress. It is repeated through the announcement as quiet proof that user funds are "more than fully backed."
I have audited enough reserve attestations to know a ratio is only as honest as its denominator. Here the denominator is user Bitcoin liabilities of 4,106.57 BTC against total holdings of 12,202.13 BTC. When a reserve ratio climbs while the disclosed liability shrinks, you are not always watching safety increase. You are sometimes watching users leave. The ledger remembers what the headline forgets.
MEXC has operated since 2018. It runs a centralized exchange positioning itself on zero trading fees and a long tail of early listings — new tokens, low-cap projects, assets that larger venues will not touch on day one. In the years since the FTX collapse, monthly Proof of Reserves disclosure has moved from differentiator to table stakes. Every major exchange now publishes some version of it. MEXC's September report is the latest entry in a genre that has quietly standardized itself: a Merkle Tree, a third-party auditor, a monthly snapshot, a press release.
The audited facts are narrow and worth stating plainly. Hacken covered four things — Proof of Liabilities, Proof of Ownership, reserve computation, and a Proof of Reserves evaluation. Four asset classes are disclosed: BTC, USDT, USDC, and ETH. Their reserve ratios land between 111% and 297%. The company's futures insurance fund is listed at roughly 798 million USDT, and a separate Guardian Fund is reported at 101 million dollars, with a stated plan to reach 500 million within two years. CEO Vugar Usi is quoted by name. The snapshot is a single point in time.
That is the whole of the verifiable record. Everything else is narrative. And the gap between the record and the narrative is where this report deserves to be taken apart, line by line.
Start with coverage. MEXC describes itself as serving more than 170 markets, spanning crypto assets, equities, tokenized assets, and derivatives. Its reserve report covers four assets. Four. The arithmetic is not ambiguous. If a platform lists hundreds of tokens and proves reserves for four, then the overwhelming majority of its liabilities carry no proof at all. Hacken did not audit MEXC's solvency. Hacken audited a scope. That scope was defined by MEXC. An auditor only vouches for the perimeter it was handed.
This is the most consequential technical flaw in the document, and it is structural rather than malicious. Proof of Reserves has always been a partial instrument dressed as a complete one. A Merkle Tree proves that a given user's balance is included in a committed set. It does not prove that the set contains every liability. It cannot. To prove completeness you would need the exchange to disclose total obligations across every asset — something no venue does, because the number would be enormous and the transparency asymmetric. MEXC discloses what it chooses to disclose. The comfortable ratios apply only to that choice.
The ratios themselves reward a closer read. USDT carries a reserve ratio of roughly 111%. USDC sits near 119%. ETH near 119%. Bitcoin sits at 297%. That spread is not accidental, and it is not a compliment. It is a map of where the platform's exposure actually lives. USDT user liabilities total approximately 1.53 billion — by far the largest single line in the disclosed book. When the dominant liability is a stablecoin at 111% backing, the real buffer against an extreme event is eleven percent, not two hundred and ninety-seven.
The 297% number demands its own audit. A reserve ratio is holdings divided by user liabilities. If user BTC deposits fall, the denominator contracts and the ratio expands without any asset being added. I ran this pattern during my 2020 review of yield aggregators — reported returns rising precisely because capital was fleeing, the headline flattering a shrinking base. The same mechanical illusion is available here. The movement from 288% to 297% cannot be interpreted without one number MEXC does not publish: the month-over-month trend in user Bitcoin balances. A rising ratio on a falling liability is a churn signal dressed as a safety signal.
The holdings themselves tell a second story. Total BTC of 12,202.13 minus user liabilities of 4,106.57 leaves roughly 8,096 BTC of platform-owned Bitcoin. That is company capital, not user coverage. It increases the ratio. It does not increase the safety of any individual depositor's claim. Presenting it inside the same ratio conflates corporate treasury with customer backing — a distinction the report never draws, because drawing it would deflate the most quotable figure in the document.
The snapshot is the next soft seam. September 10 is the anchor. September 15 is the publication date. Between those five days, and after them, the balance sheet moves. Historically, several exchanges have faced questions about temporary asset relocation around attestation dates — window dressing, in the forensic literature. A single point-in-time snapshot cannot rule this out. Only continuous, time-series reserve verification can, and that requires infrastructure no venue has shipped at scale. Silence in the code speaks louder than the pitch. The absence of any cross-time continuity proof is not an oversight. It is the boundary of what Merkle-tree attestation can ever claim.
Then there is the stablecoin concentration. MEXC's disclosed liabilities are dominated by USDT, and its disclosed reserves lean heavily on USDT as well. This creates a correlated exposure that no reserve ratio expresses: both sides of the balance sheet degrade together if the stablecoin's peg falters. A reserve ratio measures quantity on a given day. It does not measure the credit quality of the thing being counted. In 2022, reconstructing the UST de-pegging transaction flow, I watched the same structural error repeat — a system that assumed the asset on its books would hold value, priced against a liability denominated in itself. MEXC is not Terra. But it has built a solvency story on the assumption that USDT is constant. That assumption is a position, not a fact.
The insurance and Guardian Fund numbers deserve the same cold treatment. A 798 million USDT futures insurance fund sounds substantial until it is placed beside a 1.53 billion USDT user liability on a single asset. Its capacity to absorb an extreme liquidation cascade is bounded, and the report does not model that boundary. The Guardian Fund's expansion from 101 million to 500 million over two years is an unsecured forward-looking statement — no custodian named, no funding source disclosed, no third-party verification of the commitment. Forward guidance without a trustee is a sentence, not a guarantee.
The offshore dimension is the quiet one. The release originates from Mutsamudu, Comoros — a jurisdiction with no mainstream digital-asset regulatory framework. Registration and publication venues are chosen, not accidental. The report is silent on KYC architecture, licensing, sanctions compliance, and the identity of the operating legal entity. A document that devotes its length to cryptographic attestation and its margins to regulatory ambiguity has made a decision about what the reader should worry about. The map is not the territory; the chain is both. And the map here stops at the four assets.
The governance layer is thin. CEO Vugar Usi is named and quoted, which removes the most extreme anonymity risk. But his statements follow the standard trust-restoration template — references to an industry whose confidence has been "tested repeatedly," a claim that reliability "cannot be optional." The language is professionally optimized. It contains no verifiable commitment, which is precisely what a template is designed to avoid. Team backgrounds, core member histories, and investor structures are absent. For a venue operating eight years, that is a low disclosure standard.
Now to the part the bulls got right, because the forensic reading is not the same as a bearish one.
MEXC is not a house of cards awaiting a single match. It has run since 2018. It has procured a named third-party auditor in Hacken, an established Estonian-headquartered security firm, and it discloses on a monthly cadence rather than once. Continuous disclosure — even partial disclosure — is a behavioral signal, and behavior is harder to fake than a single statement. An exchange that subjects itself to repeated external scope review is doing something an exchange that never does cannot claim.
The auditor's independence has real value even inside a narrow scope. Hacken's Proof of Ownership audit at least establishes that the disclosed holdings exist and are controlled. That is not trivial. And the very existence of a monthly ritual creates a paper trail — a historical record that, over many months, becomes difficult to sustain dishonestly. History is not written; it is indexed. The value of MEXC's report is not this month's ratio. It is the accumulating series of them, which future forensic work can cross-reference against on-chain flows. A skeptic should not dismiss the record. A skeptic should insist it be completed.
The honest conclusion is not that MEXC is insolvent. It is that the report cannot tell you whether it is, and does not pretend to. Four assets against 170 markets is the load-bearing fact. The 297% Bitcoin ratio is the brightest number in the document and the one most likely to mislead, because it sits beside 111% stablecoin coverage that reflects the real cushion. The company capital inside that ratio is not customer protection. The snapshot is not a solvency guarantee. The offshore structure is not a neutral detail.
So the question that remains is not whether MEXC passed an audit. It is what a reader should do when a report proves its narrow scope flawlessly and says nothing about everything outside it. The comfortable answer is to file the report as reassurance. The correct answer is to file it as one partial input, weighted against the coverage it declines to provide — and to watch, month by month, whether MEXC widens the perimeter it is willing to have audited. Precision is the only apology the chain accepts. For now, we have precision about four assets, and silence about the rest.