A name on TON just cleared $125,000. Not a token, not a yield vault, not a protocol upgrade โ a string of characters minted as an NFT, traded on-chain where any observer can verify the price and the counterparties. The number is real, and it is already doing marketing work. But the price is not the story.
The story is one design decision that most coverage walked straight past. The achievements attached to that name are soulbound tokens โ SBTs โ the class of credential engineered to be non-transferable. Soulbound means it does not move. That is the single property separating an SBT from an ordinary token. Yet here, the entire achievement set travels when the name is sold. The buyer inherits the history. The seller keeps the money.
That contradiction is not a flaw in the pitch. It is the pitch. And when a product's core value proposition is a semantic violation, the audit checklist writes itself.
@games launched Games Names, a decentralized naming service on TON โ The Open Network, the blockchain Telegram originally incubated and later handed to an open community. Every @user-style handle is minted as an NFT. Names start at 1 GRAM with no renewal fee, minted through @games on Telegram or via the id.game entry point.
On its own, a naming service is a solved and frankly dull category. ENS on Ethereum, TON DNS on TON, Unstoppable Domains across chains, and Telegram's own Fragment-auctioned usernames all occupy the same ground. I have traced the mint and resolution paths of most of them. They differ in registry design, gas mechanics, and renewal policy โ not in concept. A name maps to an address, the address resolves, the user pays. The primitive has been stable for years.
Games Names bolts on one extra layer: games inside the @games ecosystem and its partner projects can issue SBTs to a name. Achievements, credentials, progression โ written as non-transferable tokens. The stated goal is a portable game identity, a record that follows the player across titles instead of dying inside one studio's private database. That framing is coherent, even attractive. The problem is the anchor point.
Traditional SBTs bind to a wallet address. If you sell the wallet, the credentials stay with the address. They do not follow a new owner, because the credential is a statement about who earned it, not about who currently controls a key. Games Names binds achievements to the NFT name instead. The credential becomes a statement about the name. And the name is explicitly for sale.
The roadmap is broader than the shipped product: Games Names, decentralized player profiles, first-party and partner games, a game catalog, player rankings. Only Games Names has actually launched. The rest is intent. Ecosystem position matters here too. Games Names sits at the identity layer of the TON-plus-Telegram game stack. Upstream it depends on TON for consensus and NFT standards and on Telegram for distribution. Downstream, its value depends on first-party @games titles and partner games integrating the name and issuing SBTs to it. The dependency graph runs one way: without enough games reading and writing to names, the names and their achievements have no use case. Partner-game count, integration depth, and real retention are all undisclosed. The platform's logo, designed by Art. Lebedev Studio and featured in the studio's portfolio, signals brand investment but says nothing about technical delivery.
Let me be precise about the mechanism, because the marketing deliberately blurs it.
In the standard SBT model, the token's owner is an address, and the contract enforces non-transferability at the token level. You cannot call a transfer function and move it. The credential is welded to a key pair. This is the design Buterin, Weyl, and Ohlhaver sketched in the Decentralized Society paper: reputation that cannot be bought, only earned. The guarantee is structural, not moral.
Games Names relocates the anchor. The name NFT is the owner of record. Achievements accrue to the name. When the name transfers, the achievements stay attached to it. Individual achievements cannot be sold separately โ you cannot list "Level 40 Raid Clear" on a marketplace. But the bundle moves intact. Sell the name, sell the rรฉsumรฉ.
So the real constraint is not "non-transferable achievement." It is "non-separable achievement." That distinction is the entire product, and it is also where the trust assumptions quietly shift.
The first-order economic consequence is that a transferable reputation bundle has a market price. If rankings, profiles, and matchmaking in @games titles read from the name's SBT history, then a buyer with enough capital can acquire a top-tier record without ever playing. The credential still verifies โ the chain does not lie about which name holds it. It simply stops verifying that the holder earned it. Verification and provenance diverge, and most users will not notice the gap.
I have seen this failure mode in miniature. During my 2021 Axie Infinity contract forensics, the breeding fee calculation had an edge case that allowed near-infinite token generation under specific inputs. The bug was not in the arithmetic. It was in the assumption that users would behave within intended bounds. Reputation systems share that shape. A contract can enforce that a token does not move. It cannot enforce that a name is not sold, and it cannot enforce that the buyer deserves what the name carries.
Now the supply side, which the source material leaves almost entirely blank. Name supply is undisclosed. Total count, cap, and scarcity curve are unknown. Names start at 1 GRAM with no renewal, which means no recurring revenue from the primary market โ only mint fees and whatever secondary royalties exist, also undisclosed. In my Uniswap V2 deconstruction, the fee model was legible because the invariant forced it. The product formula xยทy=k pins the 0.3% cut as a parameter inside the curve. Here there is no invariant, no curve, no published parameter. The AMM model hides its truth in the invariant; a naming protocol hides its truth in the registry and the fee schedule. Both are missing from the public record.
The GRAM unit is its own red flag. It is undefined in the source material. Historically, Telegram planned a token called Gram and shelved it after SEC action. If GRAM here is an @games-issued token, the economics change entirely and securities questions enter the frame. If it is Toncoin and the copy is merely imprecise, that still tells you something about disclosure discipline. An unclarified unit of account inside a pricing statement is not a rounding error. It is a hole where valuation should be.
Let me state the mechanism as an invariant I can actually check:
A name is an NFT with a transfer function. Achievements are SBTs whose owner-of-record is the name, not a wallet. Transfer of the name does not sever the achievements. Individual achievements have no transfer function.
The fourth line is the only thing keeping the system honest. If any single achievement becomes individually transferable โ directly or through a wrapper that lists an achievement the way a wrapped token lists an LP position โ the "earned reputation" claim collapses into an open market for credentials. So the audit question is not whether the name can be sold. It is whether the SBT contract's transfer guard is real. And the source material discloses no contract address, no audit, no open-source repository, no standard compatibility statement. Without those, the transfer guard is a promise, not a property.
The competitive picture is crowded, and the source material does not help. TON DNS is the ecosystem-native registry and likely more general-purpose. Telegram's Fragment-auctioned usernames carry the strongest brand and traffic of any handle market on the planet, because they are tied to the messaging app itself. ENS and Unstoppable Domains dominate multi-chain naming with mature tooling. Against that field, Games Names' only differentiator is the game-achievement SBT bound to a transferable name. If it cannot interoperate with the broader TON identity system, its addressable market shrinks to the @games catalog โ and that catalog is still mostly a roadmap.
Regulatory exposure deserves a mention the source material skips. Run the name NFT through the Howey factors and the picture is mixed. There is money invested โ names start at 1 GRAM. There may be a common enterprise, since name value depends on @games' ecosystem development. Expected profit depends on intent: if buyers are minting to resell or to farm a future airdrop, the fourth prong โ reliance on the efforts of others โ sharpens considerably. Name NFTs used purely as in-game identity and collectibles sit at lower risk. Names bundled with token rewards or profit expectations edge toward an investment contract. And @user-style handles invite trademark squatting, impersonation, and brand-name capture once a public secondary market exists โ a compliance headache with no disclosed KYC, AML, or legal structure to manage it.
Zero knowledge isn't magic; it's math you can verify. The same discipline applies to reputation. I don't care what the marketing says about soulbound credentials. I care whether the contract exposes a path to move a single achievement, and whether anyone independent has confirmed it does not.
The consensus read on Games Names is favorable: a clever composition of TON, Telegram distribution, GameFi, NFT naming, and SBTs โ five live narratives in one product. That is precisely the packaging I distrust. When a launch stitches together every hot sector, the technical substance is usually the thinnest layer, because the narrative does the work.
The contrarian point is colder. Transferable SBTs do not extend soulbound identity. They invert it. The value of an SBT comes from the guarantee that the holder is the earner. The moment the credential rides a tradeable container, you have built a market for identities โ and markets for identities attract exactly the wrong buyers: account farmers, leaderboard manipulators, and anyone who wants a rรฉsumรฉ they did not build.
The comparison that matters is not ENS versus Games Names on registry design. It is whether a game's ranking means anything if the top slot is purchasable. Galxe and similar credential platforms already bind to wallets and face the same farming problem โ but they keep the credential on the earner's key. Games Names removes even that friction. You do not need to compromise a wallet or phish a seed phrase. You need a transfer and a payment.
There is a second blind spot: the silence itself. No audit, no open source, no contract standard, no supply disclosure. In a bull market, silence reads as momentum. In an audit, silence reads as unverified attack surface. Every project I have dissected with a wide roadmap and a thin technical disclosure hid its risk in the gap between narrative and code. The 2018 Gnosis Safe work taught me this from the other direction: three signature malleability issues survived early review because auditors trusted the interface instead of reading the ECDSA handling. Trust the interface and you miss the bug. Trust the roadmap and you miss the gap.
During my 2024 review of institutional custody designs ahead of the spot Ethereum ETF, the distance between the marketing deck and the actual multisig architecture was the whole story. Threshold schemes looked decentralized on the slide and centralized in the configuration. Games Names has the same shape: a decentralized-naming label layered over a platform-controlled issuance model, where @games and its partners decide who receives which credential.
The reputational risk is not theoretical. If a name with a strong SBT history sells, the buyer's on-chain record is indistinguishable from a legitimate veteran's. Ranking systems, tournament eligibility, and partner-game rewards that read from that record become purchasable. The protocol cannot detect it, because the transfer was legal by its own rules. That is the defining trait of a design flaw rather than an exploit: no one has to break anything.
Watch three signals, not the price of a name.
First, does @games restrict high-value achievements from transferring with the name โ a non-portable allowlist, or a re-verification step on transfer? If not, the leaderboard is for sale.
Second, does an audit and a contract address appear? Until they do, the transfer guard is unverified and the "soulbound" label is marketing.
Third, what does "1 GRAM" actually mean, and is name supply capped? Those two answers decide whether this is an identity layer or an unbacked collectible.
My forecast: the transferable-reputation model gets copied before it gets audited. It is too convenient a growth loop to ignore โ names with history command a premium, premiums drive minting, minting drives the narrative. The ethics arrive later, if at all. The real question is not whether Games Names works. It is whether a soulbound token you can sell is still soulbound โ or just a credential with a price tag attached.