Reading the room in a room of code, the loudest thing this month wasn't a token launch or a rollup upgrade. It was a sentence buried in a service agreement.
Sony told its users, in the plainest legal language available, that the games they buy are not theirs. Not the disc images, not the entanglements, not the digital entitlements sitting in their account libraries. What they hold is a permission slip โ revocable, non-transferable, and subject to change at the vendor's discretion.
Within hours, the crypto timeline had a rebuttal ready. NFTs, the argument went, can deliver real digital ownership. Immutable ledgers. Self-custody. No one can revoke what lives on-chain.
I've watched this exchange repeat itself for five years now, in slightly different costumes. And this time I want to slow down, because both sides are answering a question that isn't the one being asked.
The Narrative Cycle Nobody Notices Is Cycling
The digital ownership debate has a documented rhythm. In 2017, CryptoKitties proved a game asset could exist independently of the game's server โ briefly, until the contract congested Ethereum and the cats stopped breeding. In 2021, the PFP wave reframed NFTs as identity rather than utility. I spent that spring interviewing collectors who described their avatars as membership cards rather than JPEGs, and I predicted the shift from "asset" to "access key" months before the market corrected. In 2022, floors collapsed, and the industry learned a distinction it should have learned in 2017: asset price and narrative value are not the same variable.
What's changed by 2026 is that the institutional layer has arrived. Sony Interactive Entertainment operates Soneium, an Ethereum layer-2 built with Startale, alongside a patent portfolio covering cross-platform asset transfer. Nintendo runs its own closed economy. Microsoft has spent a decade licensing its way through every digital storefront dispute on record. The entities now discussing digital ownership are not startups with whitepapers. They are platform holders with balance sheets, legal departments, and a demonstrated preference for controlling the terms of trade.
So when Sony says "you don't own your games," it is not making a philosophical statement. It is making a commercial one, vetted by counsel, consistent with the licensing model that has underwritten the console business since the first EULA was printed. The interesting question is not whether that's fair. The interesting question is what an on-chain token can actually change about it.
Core: Three Problems, Only One of Which Is a Token Problem
Let me get concrete, because abstraction is where this debate usually dies.
When a player buys a game, the legal instrument is a license, not a title. The distinction is old and unglamorous: title conveys ownership of a thing, license conveys permission to use a thing under conditions. The vendor retains the thing. The buyer retains the conditions. Under this structure, "buying" a game is closer to renting a hotel room with unlimited nights than to purchasing a house.
An NFT does not change this. What it changes is the property character of the receipt.
Here's the mechanism. A non-fungible token is a ledger entry โ a unique identifier with an owner field, governed by a standard like ERC-721 or ERC-1155. It is extremely good at proving that address A holds credential B, and at transferring that credential to address C without asking permission from a central registry. What it is not good at is enforcing what credential B entitles you to inside a game whose rendering, physics, anti-cheat, and economy all run on a vendor-controlled server.
This is the first real problem, and it isn't a blockchain problem. It's a three-way consistency problem.
The three parties are the client (installed on the player's machine), the server (authoritative game state, owned by the vendor), and the chain (the ownership record). For an NFT to mean anything in-game, all three must agree in real time. The client must recognize the token and render the right item. The server must trust the client's claim without opening itself to injection attacks. The chain must reflect state changes without forcing a transaction confirmation into the frame budget of a 60fps shooter.
No production system has solved this at scale. The closest analogues โ Ubisoft's Quartz, various GameFi experiments, assorted SDK startups โ have all shipped half of the pipe and hoped the other half would appear. It didn't. What's missing isn't a token standard; it's a cross-industry interoperability protocol, and no single company has the incentive to build one alone.
A note on the infrastructure layer, since Soneium is an L2 and everyone assumes that's the relevant fact. It isn't. The data availability question โ how much compressed state a rollup posts back to Ethereum โ gets treated as the central design constraint of every L2, and for the vast majority of them it is a solution in search of a problem. Very few rollups generate enough data to saturate a dedicated DA layer, let alone justify the cost of one. Soneium's asset registry, if it ships, will carry a volume of writes comparable to a mid-sized regional marketplace. That is a database with a hash pointer, not a data-availability scaling problem. The genuinely hard engineering in game assets sits in client-side verification and server-side trust minimization. Nobody discusses that, because it doesn't produce a tradable token.
The Privacy Gap Has Two Meanings, and They Are Not Equally Bad
The second problem is the privacy gap, and here I want to be precise, because the language is ambiguous in ways that matter. "Privacy gap" could mean two very different things, and the risk magnitudes are not comparable.
Interpretation one: on-chain assets are empirically linkable. Wallet addresses are pseudonymous, not anonymous. Chain-analysis firms built an entire discipline around this fact. If your game library lives on-chain, so does your spending history, your holding period, your counterparties, and the timing of every purchase โ derivable by anyone with a graph database and patience. This is a technical problem with technical mitigations: zero-knowledge proofs, privacy addresses, stealth transfers, or a custodial layer that defeats the purpose. I spent my final undergraduate year at Tartu verifying Zcash's proving system in Python, line by line, because I wanted to know whether the cryptography actually held. It does, in narrow deployments. At game-scale throughput, with millions of low-value transactions, the cost-benefit math gets ugly fast. Zero-knowledge proofs are cheap to verify and expensive to generate, and a game that mints a token every time someone picks up a sword is not a game that can afford a proving circuit per sword.
Interpretation two: on-chain immutability conflicts with data protection law. This is the one that keeps compliance teams awake. The right to erasure requires that personal data be deletable on request. A public blockchain is designed to make data permanent. These two requirements are not reconcilable by engineering โ you either keep the data off-chain (and lose the composability that made the asset interesting) or you accept that a European regulator may eventually classify your ledger as non-compliant infrastructure. Interpretation one is a cost problem. Interpretation two is an architecture problem that may have no solution inside the current legal framework. The source material doesn't tell us which one is intended, and that ambiguity matters enormously, because one is solvable and the other might not be.
There's a third issue neither side likes to discuss, because it embarrasses both: transfer rights and usage rights are separable.
When an NFT is sold, what moves is the credential and the transfer right. What stays put is the usage right โ the ability to actually play with the item inside a working game. If the game shuts down, the token remains and the utility goes to zero. The NFT becomes, in the industry's favorite black humor, a JPEG with provenance.
This is not a bug that better engineering fixes. It's a structural consequence of where utility lives. Utility lives on the vendor's server. The token lives on the chain. The token cannot compel the server to keep running, and the server has no obligation to honor a token it didn't issue. Ownership of the receipt is not ownership of the thing, and the gap between them is exactly the size of the vendor's ongoing cooperation.
There's a governance parallel I find hard to ignore, because I've spent enough nights inside DAO forums to recognize the pattern. On-chain governance routinely reports voter turnout below 5% of token supply. The votes that do land are dominated by a handful of addresses โ foundations, early funds, and a rotating cast of delegates whose mandates are, at best, loosely defined. "Community decision-making" is an aspiration, not a description. The same asymmetry applies to game assets: the community can hold the token, but the platform holds the settlement layer, the rendering layer, and the terms of service. Decentralization at the credential layer with centralization at every layer that matters is not a new structure. It is the old structure with better bookkeeping.
The Contrarian Read: It's Motivation, Not Technology
Here's the part the crypto side gets wrong, and I say this as someone who has spent six years arguing that on-chain credentials are genuinely useful.
The obstacles to digital ownership are not primarily technical. They are motivational.
Consider what a platform holder surrenders by allowing true asset portability. It gives up the ability to reprice its catalog unilaterally. It gives up the second-hand market's cut, or at least the uncontested version of it. It gives up the ability to sunset a title and let the associated assets quietly evaporate. It gives up the ability to rebalance an in-game economy when inflation runs hot, because the supply of a given item is now externally verifiable and externally owned.
Every one of these is a revenue lever, not a technical inconvenience. No vendor surrenders revenue levers because a better token standard appeared. The history of the digital storefront is a history of vendors resisting exactly this โ and prevailing.
This is the innovator's dilemma in its purest form, and the historical resolution is not that incumbents adopt the new model. It's that new entrants start where the incumbents cannot afford to. CryptoKitties didn't come from Nintendo. Axie didn't come from Blizzard. The open-standard version of game asset ownership โ if it arrives โ will arrive from studios with nothing to protect, not from platform holders with everything to lose.
Which brings me to the sentence I keep circling. The claim is framed as something "crypto advocates believe." Read that construction carefully. When a position is reported as what its adherents believe rather than as what the evidence shows, the language itself is a temperature reading โ the narrative is in its argumentative phase, not its delivery phase. In 2021, articles said "NFTs are doing X." In 2026, they say "advocates argue NFTs could do X." The tense shifted because the receipts didn't arrive.
Sony, meanwhile, is not standing still. A company that runs its own Ethereum layer-2 while telling users they don't own their purchases is not rejecting on-chain assets. It is positioning to issue them under its own terms, on its own chain, with its own settlement rules and revocation policy. That is not a contradiction. That is a strategy. And it sits inside a broader institutional pattern worth naming: the same entities now piloting programmable currency are the ones whose asset systems require a persistent identity layer. A token that can be transferred without permission and a ledger that records every transfer are not complementary designs. They are opposites, and the architecture will eventually force a choice.
Takeaway: Watch the Contract, Not the Chain
I don't think the next twelve months will resolve any of this. I think they will clarify which layer the fight is actually on.
Watch the contract language, not the whitepapers. Watch whether platform holders publish an asset-export policy. Watch the digital content disclosure rules moving through European and American legislatures, because a legal requirement to tell consumers they are buying a license rather than a thing will do more for transparency in digital ownership than any token standard shipped this cycle. And watch second-hand market volumes during the dead weeks, when no incentive program is running, because that is the only number that distinguishes demand from subsidy.
The chain can prove who holds the receipt. It has never been able to prove who holds the keys.