A new security assessment of Tether’s USDT has surfaced a structural risk that has less to do with hackers attacking blockchains and more to do with how few keys stand between them and nearly half the world’s largest stablecoin supply.
Cybersecurity firm Hacken this week published a hands-on assessment of USDT across the TRON, Ethereum, and Solana blockchains — the three networks that together account for roughly 98 percent of its native supply, totaling about 184.6 billion USD. The headline finding: about 91.3 billion USD of USDT on TRON is governed by a smart contract controlled through a 2-of-3 multisig arrangement. Compromise just two signing keys, and an attacker could gain privileged control over actions affecting that entire pool.
## Two keys, ninety-one billion dollars
The TRON deployment is the single largest concentration of USDT anywhere — roughly half of the total supply. Privileged administrative transactions on that contract require approval from two of three designated key holders. Hacken’s analysts found no way to seize administrative control with a single key, which is reassuring as far as it goes. But the 2-of-3 threshold means the security of nearly half of all USDT ultimately rests on the integrity of three signatures.
The report, which fed into Hacken’s bluechip rating for the token, also flagged the absence of safety infrastructure that decentralized finance projects have treated as standard for years. The contract has no timelock, no cancellation mechanism, and no other way to roll back administrative changes once the required signatures are collected. Minting new tokens, freezing addresses, and — where applicable — upgrading smart contracts all take effect immediately.
A timelock, Hacken argued, would create a window between approval and execution during which defenders could detect and stop a malicious operation before it lands. For a 184 billion USD asset, there is currently no such window.
## Key reuse multiplies the blast radius
The concentration problem compounds across networks. Hacken found that the same set of signing keys is reused across multiple blockchains: Ethereum, Avalanche, and Celo all use the same six keys under a 3-of-6 approval scheme. On Avalanche and Celo, those keys additionally control the ability to replace the token’s code entirely.
That means a single compromise event — three signers, one shared key set — could potentially impact multiple USDT deployments at once, not just one network. In security terms, the effective attack surface is smaller than the number of blockchains suggests.
## Issuance without onchain limits
Hacken’s review also examined USDT’s proof of reserves and how it connects — or fails to connect — to the mechanics of issuance. The firm factored the reserves reporting into its assessment but lowered its rating because there is no onchain link between reserve verification and the actual minting of tokens.
In plain terms: after a transaction is authorized by the required signer quorum, the audited contracts impose no additional limits on how much USDT can be issued. The onchain guardrails that decentralized protocols often build into minting paths — caps, oracle checks, reserve attestation hooks — are absent from USDT’s architecture, by design. Tether’s model has always relied on corporate process and attestation rather than smart-constraint, and the report makes the trade-off explicit.
## Why it matters beyond Tether
The findings land at a moment when stablecoins are becoming financial market infrastructure rather than crypto curiosities. USDT is the most widely used trading and settlement asset in the industry, and its supply has grown to a scale where a governance failure would not stay contained inside crypto markets. Regulators in the United States, Europe, and Asia have spent the past two years drafting stablecoin regimes that focus heavily on reserves and redemption rights; operational key management has received far less attention.
Hacken’s assessment is a useful corrective. Reserves answer whether an issuer can honor redemptions. Key governance answers whether an attacker can rewrite the rules before anyone can react. The 2-of-3 multisig guarding 91.3 billion USD on TRON, with no timelock and no rollback path, is the kind of single point of failure that the industry would roundly criticize in any other protocol.
None of this is a claim that USDT is about to be exploited. Hacken explicitly found single-key compromise insufficient for seizing control, and Tether’s operational security has improved markedly since the industry’s early years. But the report quantifies, for the first time in a formal security review, exactly how concentrated the stablecoin’s administrative power is — and how little reaction time the market would have if the wrong two keys ever fell into the wrong hands.
Price snapshot at publication: BTC traded around 79,000 USD, ETH near 2,479 USD, and SOL around 104 USD (14:45 UTC, Sept. 7, 2026).
98 percent of USDT sitting on three chains and the biggest chunk answers to three keys. every audit season we relearn this and nothing ships
91.3 billion USD behind a 2-of-3 multisig is wild. Two compromised keys and half of all USDT is at the mercy of whoever holds admin. Hacken did the industry a favor putting a number on this.
The part people miss: a single key gives nothing, so people assume it is fine. But nation state level attackers do not need one key, they need two. Social engineering two key holders is very plausible.
Bergljot exactly, and the key holders are humans with laptops and phones, not cold vaults in a mountain. The 98 percent concentration on three chains makes it worse.
2-of-3 multisig for 91 billion. a random defi fork launched last week ships with a 48h timelock and this contract doesnt. absurd
A 48h timelock would at least let everyone react before admin keys touch anything. Tether treats the TRON deployment like a side project and it holds half the supply.
To be fair, Hacken found no single-key takeover path. The real risk is key custodianship, not the contract logic itself.
custodianship IS the contract risk though. Hacken can test code paths all day, they cant test whether two of three humans get phished next quarter
Tether should move to a higher threshold multisig or timelocked admin on TRON. The tech exists, this is a governance choice at this point.