A namespace becomes an asset the moment it can leave its owner, and platform usernames cannot. WhatsApp handles, X rare handles, and Fragment usernames are all priced as if they were property, but none of them can be sold to an unrelated buyer, licensed to a subsidiary, or pledged against a facility — which means they fail the single test that made domain names durable IP.

The claim worth testing

The prevailing read on 2026’s username markets is that handles have become a new asset class, and the price tags support it. Rare handles on X start at $2,500 and reach seven figures, according to X’s marketplace documentation (2026). Fragment, Telegram’s official marketplace, has cleared premium @Names well into six figures.

Price is not the test, though. Fine art, private company stock, and airline miles all carry prices. What separates an asset from a balance on someone else’s ledger is whether the holder can move it — and it is precisely that property the platforms have withheld.

Liquidity is a property of the market, not the price tag

Three conditions make a namespace genuinely tradable: the holder can transfer it to an unrelated party, the asset survives the holder leaving the platform, and a third party can value it without the platform’s cooperation. A handle that fails any one of these is a subscription, not a holding.

WhatsApp’s implementation, rolled out from 7 July 2026, is first-come-first-served and non-transferable: the handle is reserved to an account, and when the account closes the handle returns to the pool. X prohibits handle flipping outright, so a buyer who later ends their subscription loses both the handle and what they paid for it. Fragment comes closest — the username is a token on a public ledger and can genuinely be sold — but what the buyer receives is a Telegram identity, usable nowhere else.

What the domain aftermarket actually proved

The domain market is the standing counter-example, and it is worth being precise about why. The reason AI.com changed hands for roughly $70 million, according to NameBio (July 2026), was not scarcity — plenty of scarce things do not trade. It was that the buyer could move the name to any registrar, host it anywhere, park it, license it to a subsidiary, or resell it to someone with no relationship to the seller.

That portability is what a valuer, a lender, and an acquirer’s counsel each need before a name can appear as anything on a balance sheet. Twenty years of resale history did not create the value; transferability did, and the resale history is simply the evidence.

The captured-pool problem

Every platform username market built so far is a captured pool by design. Value accrues inside it, and the platform sets the terms on which any of it may leave — usually, not at all. That is a rational choice for the platform: a handle that cannot leave is a retention mechanism, and the revenue from selling access to it is recurring rather than one-off.

For the holder it produces a peculiar exposure. A business can spend years making @companyname the address customers actually use, and hold nothing it can sell, insure at replacement cost, or carry through an acquisition. If the feature is deprecated — and platform features are deprecated routinely — the accumulated recognition does not transfer anywhere. It simply stops.

Treat the handle as a pointer, not the property

The workable position is to let platform handles do what they are good at and nothing more. They are excellent discovery and routing surfaces: cheap to claim, easy to publish, and they shorten the path between a customer and a conversation.

They are poor foundations. Authentication, customer records, and the identity a business actually owns belong on a domain, which survives any single platform’s product decisions. Claim the handles defensively, point them at infrastructure you control, and do not let a namespace you cannot sell become the thing your customers know you by.

The clearest signal of how this is already understood will come from deal documents. Watch for an acquisition where the target’s handles are explicitly excluded from the assets acquired — that exclusion is a valuation, and the number it implies is zero.

Key takeaways

  • A namespace is tradable only if it can be transferred to an unrelated party, survives the holder leaving the platform, and can be valued without the platform’s cooperation; platform usernames fail at least one of these tests in every current implementation.
  • WhatsApp handles are non-transferable and return to the pool when an account closes; X prohibits flipping and ties continued access to an active subscription, so a lapsed subscriber forfeits both the handle and the purchase price.
  • Fragment usernames are genuinely transferable on a public ledger, but what transfers is a Telegram identity rather than a namespace usable on any other service.
  • The domain aftermarket’s durability rests on portability, not scarcity — the roughly $70 million paid for AI.com, according to NameBio (July 2026), bought an asset the purchaser could move, license, or resell to anyone.
  • The defensible posture is to treat handles as discovery pointers to domain-based identity that the business actually controls, and to expect handles to be written out of acquisition consideration entirely.