Justin Sun Prize: How a Math Award With On-Chain USDT Payouts on TRON Works

The Justin Sun Prize is an award where the payout decision is made not by a jury but by a verification program. TRON founder Justin Sun announced an initiative in which the prize fund has been moved on-chain in advance, the wallet address has been published so anyone can check the balance, and the reward goes to the recipient once a proof has been checked by a machine "from the first line to the last." Payouts run through USDT TRC-20 on TRON or USDC on Ethereum, the acceptance criterion is machine formal verification (in the cases described, formalization was done in Lean), and the most unusual part of the design is precisely the on-chain payouts: according to the organizers, the funds cannot be withdrawn and a problem cannot be removed from the registry. Let's look at how this works technically and what questions the scheme leaves open.

What was announced and when

Justin Sun's office announced the prize on September 16, 2026, in Geneva, describing it as a decentralized academic bounty mechanism built around a "problem list" with machine formal verification and a maximum award of up to $1 million per problem. The list of problems, the confirmation standards and the verification materials are published on GitHub, payout records are logged on-chain, and the recipient chooses the network: USDT on TRON (TRC-20) or USDC on Ethereum (ERC-20).

The first list includes 66 classical mathematics problems — under the slogan "66 problems. One limit." Among the entries in the registry are the formal verification of the Poincaré conjecture, the Riemann hypothesis, the Goldbach conjecture and a whole series of Paul Erdős problems. According to NewsBTC, the $1 million top award went to an OpenAI research team for the existence and smoothness problem for the three-dimensional Navier–Stokes equations, with the Lean formalization produced by GPT-6 Astra. The prize materials are hosted on the project page. The inscription on the medal reads "Quod probatur, solvitur" — "What is proved is paid."

Sun himself describes his role modestly: "I am not a mathematician. I am merely an observer and a benefactor. The rest I leave to the proofs." According to published reports, his powers are limited to compiling the list of problems and setting the reward amounts — once a problem is added to the registry it cannot be removed, and the funds cannot be reclaimed.

The mechanics: prover, formalizer and the "awaiting formalization" status

The prize differs fundamentally from classical scientific awards: there are no age limits, nominations or annual ceremonies. Instead, there are two reward-recipient roles.

Role	What they do	Who can fill it
prover	finds the proof	a human, an AI or a combination of both
formalizer	translates the proof into a machine-checkable format	a human, an AI or a combination of both

Both roles can be filled by a single participant or by different people and systems. No barriers based on nationality or institutional affiliation have been declared, and no distinction is drawn between humans and AI.

An intermediate scenario is spelled out separately: if a proof is accepted by the community but has not yet been formalized, the problem is given the status "proved, awaiting formalization," and the funds remain locked. In other words, public recognition alone does not trigger a payout — only machine verification does.

Why machine verification rather than expert review

The "arbiter" role in this scheme goes to a formal prover not because it is fashionable but because of the properties of the tool. Lean 4 is an open-source prover and programming language that has been used to build and check large-scale proofs; its extensive mathematical library makes it convenient both for reasoning about mathematics and for analyzing DeFi contracts with complex calculations. It is in Lean, according to reports, that the award-winning proof was formalized.

The second reason is the sharp growth in the number of proofs found with the help of AI. When the author of a solution is a model, the guarantee of correctness becomes a machine-checked formalization rather than a coherent natural-language text. The scale of the problem is illustrated by statistics on the Erdős problems: according to DeepMind, out of 200 candidate solutions labeled either correct or incorrect, only 6.5% turned out to be "substantively correct." For a prize with automatic payouts this is the key argument: without a machine-checked proof free of sorry, any review turns into an endless expert debate — and into an argument over who should be paid.

The flip side is how attainable the criteria are. In those same tests, a commercial Lean agent failed to formalize two papers after dozens of hours of work, while a specialized harness covered all seven target theorems with no sorry. In other words, the formalization bar is high, and some problems may remain in "awaiting formalization" status for a long time.

On-chain payouts in USDT TRC-20: how it works

Technically, paying out a prize in USDT on TRON is not "just a transfer" but a call to the token's smart contract. USDT on the TRON network is issued under the TRC-20 standard, which ensures compatibility with protocols and dApps on TRON; the contract address is TR7NHqjeKQxGTCi8q8ZY4pL8otSzgjLj6t. All TRC tokens use at least 6 decimal places (TRX, USDT and USDC use 6, USDD uses 18), which matters when reconciling amounts in an explorer.

A few practical network parameters that shape the design of any prize scheme:

• Block time is about 3 seconds, with consensus maintained by 27 Super Representatives. A payout is recorded almost immediately after the transaction is sent.
• Energy is mandatory. Every activated account receives 600 free Bandwidth per day, but there is no free Energy for contract calls — a USDT transfer requires Energy or the burning of TRX.
• fee_limit. The current price of one unit of Energy is 100 sun (0.0001 TRX), with 1 TRX = 10⁶ sun; the maximum fee_limit value on mainnet is 15,000 TRX.
• New addresses. Activating a new account costs a fixed 1 TRX on top of the usual Bandwidth cost, and if Bandwidth is insufficient, an additional 0.1 TRX is burned.

It is worth separating marketing from the technology here: TRON's promotional pages talk about "zero-fee" USDT transfers, meaning there is no percentage charge for the transfer, while the technical documentation describes a resource model — contract calls consume Energy and Bandwidth. Formally this is not an error in the documentation but two levels of describing the same process. For a prize recipient it boils down to a simple rule: it helps to keep a little TRX on the address where the USDT will arrive (or to have access to a resource-payment mechanism) — otherwise you won't be able to move the funds.

A public wallet instead of promises

The main substantive part of the "on-chain" approach is not speed but verifiability. The prize fund was moved on-chain in advance, the address has been made public, and anyone can check the balance. That turns a statement about the fund from a press release into an observable fact, and turns the declared irreversibility of payouts and the impossibility of removing a problem from the registry into a limitation on the founder's own powers.

This is not the first on-chain bounty for proofs

The idea of paying for theorems directly from a blockchain existed earlier. In the Proofgold project, users post bounties on conjectures: as of early 2025, more than 13,000 bounties had been posted, of which nearly 3,000 were claimed — either for proving a conjecture or for refuting it. The same project uses a commit-and-reveal scheme to protect authorship, which is why publishing a document on-chain takes about a day, with blocks arriving roughly once an hour.

The comparison is telling: TRON, with roughly 3-second blocks, records a payout almost instantly, but the question of protecting authorship until the proof is revealed is addressed not by the network's properties but by the prize rules and the publication of materials on GitHub.

Risks and what users should check

Any high-profile "prize in USDT" is an obvious lure for scammers, and that is not a hypothesis but a standard scenario. TRON's documentation specifically advises verifying a token's contract address in TRONSCAN and warns about copycat tokens with similar names such as TRX2 or USDT-V2, which are almost always scams. Unsolicited airdrops to an address usually serve as bait, and sites offering to "verify your wallet" with a signature often request an approve for unlimited TRC-20 spending.

What is worth checking in substance:

1. The primary source of the announcement — Justin Sun's and TRON DAO's official channels, not second-hand retellings.
2. The fund address and payout transactions in TRONSCAN, including whether the amount matches what was announced.
3. The USDT TRC-20 contract address — a token with the "right" ticker but a different contract has nothing to do with Tether.
4. The problem statement and acceptance criteria in the repository: which prover and which library count as acceptable, the ban on sorry, and the procedure for establishing authorship.
5. Who acts as arbiter in disputed cases and in which jurisdiction the fund is based.

Conclusion

The Justin Sun Prize is interesting not for the size of its fund but for the change in procedure: instead of a committee that reads a text and votes, there is a verifier that checks a formalization and a blockchain that executes the payout. The prover and formalizer roles acknowledge a reality in which a proof may be found by an AI, while the declared irreversibility of the registry constrains the founder himself. At the same time, the scheme inherits all the accompanying limitations: a high formalization bar, dependence on how well the problems are stated, and the vulnerability of a high-profile brand to phishing. Verifiability cuts both ways here — and facts are worth confirming through the explorer and primary sources rather than headlines.

This material is for informational purposes only and does not constitute investment advice.