The lazy version of decentralized identity says users will own their data on-chain. That is both technically wrong and politically dangerous. A public blockchain is a hostile place to store identity because it is permanent, indexable, and adversarial by default. The serious version of self-sovereign credentials uses blockchains for what they are good at: timestamping, issuer registries, revocation commitments, and neutral verification rails. The passport, payroll record, age proof, or accredited-investor status should remain off-chain, encrypted, and selectively disclosed by the holder.
This distinction matters now because identity is no longer a crypto side quest. The EU Digital Identity Wallet is moving under eIDAS 2.0, MiCA is hardening exchange compliance, tokenized treasuries require qualified participant checks, and airdrop farming has turned Sybil resistance into a direct cost center. With ETH near $1,725 and SOL around $73 in the supplied market snapshot, price action is not the signal. The signal is that on-chain finance is being forced to answer a dull but decisive question: who is allowed to do what, without rebuilding Web2 surveillance?
The Consensus View Is Backwards
Most identity-token projects start with a consumer narrative: mint a soulbound token, build reputation, and carry it across DeFi. That framing confuses recognizability with identity. A soulbound token saying an address passed KYC at an exchange may be useful for a sanctions-screened pool, but it is also a correlation beacon. Once attached to a trading wallet, it links lending behavior, governance votes, NFT purchases, bridge history, and liquidation events to the same compliance marker.
The better architecture starts with the W3C Verifiable Credentials model: an issuer signs an attestation, a holder stores it in a wallet, and a verifier checks a proof. The blockchain enters as a minimal public coordination layer. It can anchor decentralized identifiers, publish issuer public keys, store revocation roots, and provide an immutable audit trail that a verifier can query. It should not contain the credential payload. Putting identity data on-chain is not decentralization; it is doxxing with extra gas fees.
There is a reason enterprise identity vendors, regulators, and the more serious Web3 infrastructure teams are converging on selective disclosure. BBS+ signatures, CL signatures, SD-JWT, and zero-knowledge circuits allow a user to prove a predicate without revealing the underlying record. The useful proof is not I am Alice born on this date. It is I am over 18, I am not on this revocation list, I am resident in an eligible jurisdiction, or I control a credential issued by this university. That is a fundamentally different product from an on-chain badge.
What Actually Belongs On-Chain
A production-grade self-sovereign identity stack has four on-chain components, none of which are glamorous. First is an issuer registry, where verifiers can resolve whether a credential signer is legitimate. This can be a DID document anchored to Ethereum, a permissioned registry governed by a consortium, or a smart contract controlled by a regulated entity. The hard problem is not the JSON; it is governance over key rotation, issuer suspension, and liability when an issuer is compromised.
Second is revocation. Credentials expire, fraud is discovered, licenses are suspended, and sanctions lists change. If a verifier cannot check revocation state, the credential is theater. The clean design is a Merkle root or cryptographic accumulator updated periodically on-chain, with holders producing membership or non-membership proofs. A 32-byte root published every block is wasteful; a batched root per hour or per day is usually enough for education, employment, and many financial use cases. High-frequency compliance, such as exchange withdrawals, may need faster revocation but should still avoid publishing user identifiers.
Third is proof freshness. A verifier needs confidence that a presentation was generated for a specific session, domain, and time window. This prevents replay attacks where a valid age proof or KYC proof is copied and reused. Blockchains can supply timestamps and challenge registries, but the heavier work belongs in wallets and verifier infrastructure. A wallet that cannot bind a proof to an application domain is not self-sovereign; it is a phishing surface.
Fourth is dispute and audit infrastructure. If a lending protocol grants undercollateralized credit based on credentials, somebody must be able to prove which issuer key, revocation state, and verifier policy were in force at the time. This is where chains beat traditional databases. They provide a neutral, append-only record that survives vendor failure. The credential remains private, but the verification context becomes auditable.
Zero-Knowledge Identity Is Useful, but Not Magic
ZK proofs are becoming the default marketing layer for decentralized identity, and for once the hype has a real technical base. Polygon ID, now evolving under Privado ID, demonstrated issuer-holder-verifier flows using zkSNARKs. Semaphore showed practical anonymous group membership proofs. zkTLS projects are pushing toward proofs about Web2 data, such as bank balances or exchange account status, without asking institutions to issue native credentials on day one. These are important primitives.
But ZK does not solve the two hardest identity problems: source integrity and social recovery. A proof that a credential is valid says nothing about whether the issuer performed adequate checks. If a fake university issues a mathematically pristine degree, the proof system has worked and the identity system has failed. Likewise, if a user loses the private key controlling her credential wallet, cryptography does not magically restore her professional licenses or medical records. Recovery requires institutions, guardians, hardware enclaves, or regulated custodians. Purists hate that answer; production systems need it.
The trade-off is also computational. General-purpose ZK circuits remain expensive to design, audit, and maintain. Proving simple predicates such as age over 18 or country not equal to a prohibited jurisdiction is realistic. Proving complex policies across multiple credentials, revocation states, and real-time risk scores can become brittle quickly. The practical market will favor standardized circuits, audited proving libraries, and boring policy templates over bespoke cryptographic art projects.
The winning identity stack will look less like a social graph and more like payments infrastructure: standardized messages, regulated endpoints, cryptographic receipts, and ruthless uptime requirements.
The Market Demand Is Coming From Compliance, Not Ideology
Crypto founders like to pitch decentralized identity as user liberation. Buyers are more pragmatic. Exchanges need Travel Rule workflows. DeFi protocols considering real-world assets need jurisdiction and investor-status checks. DAOs need Sybil resistance. Games need bot mitigation. Social apps need proof-of-personhood without storing government documents. The budget is attached to fraud reduction, regulatory access, and customer acquisition, not philosophical sovereignty.
Gitcoin Passport is a useful case study. Its value is not that every stamp is perfectly decentralized; it is that projects can raise the cost of Sybil attacks by aggregating signals from accounts, attestations, and social proofs. Worldcoin, despite its deserved controversy over biometric enrollment, proved there is market appetite for proof-of-personhood at scale: the World App passed 10 million users in 2024, while verified human counts reached the multi-million range. The lesson is not that iris scans are the future. The lesson is that scarce human uniqueness is economically valuable when bots can farm tokens, quests, and governance.
Coinbase and other centralized platforms are also quietly positioned to become major credential issuers. A regulated exchange can attest that a user completed KYC, resides in a jurisdiction, or meets certain account-history thresholds without revealing the user’s legal name to every DeFi protocol. That is uncomfortable for decentralization maximalists, but it is probably how the first billion compliance-grade credentials are issued: by institutions that already have customer files and regulatory obligations.
Chain Choice: Ethereum Security, L2 Economics, and the Solana Question
Identity builders obsess over which chain to use, but the correct answer depends on update frequency and trust assumptions. Ethereum mainnet is credible for high-value issuer registries and revocation commitments that change infrequently. The cost is acceptable if updates are batched. Layer-2 networks such as Base, Arbitrum, Optimism, and Polygon zkEVM are better for frequent attestations, policy registries, and application-level verification logs, especially after EIP-4844 reduced data availability costs for rollups by more than 90% in many periods.
Solana offers low-latency and cheap state updates, which is attractive for consumer identity flows, gaming credentials, and high-volume proof routing. The trade-off is different: finality assumptions, validator economics, and the weaker institutional comfort that still surrounds Ethereum as the settlement layer for regulated assets. If a tokenized treasury fund needs credential-gated transfers, its lawyers will likely prefer Ethereum or an Ethereum-aligned L2. If a game needs 5 million bot-resistance checks per day, Solana’s performance profile becomes harder to dismiss.
The multichain answer is not to replicate identity state everywhere. That creates fragmentation and inconsistent revocation. A better pattern is a canonical issuer and revocation registry on a settlement chain, with proofs portable across applications through standards. Cross-chain messaging can mirror roots, but verifiers should understand which root is authoritative and what finality delay applies. Identity is too sensitive for bridge theater.
Design Rules for Builders and Investors
The strongest decentralized identity projects will share several traits. They will minimize on-chain personal data to zero. They will support open standards rather than proprietary wallet lock-in. They will treat revocation as a first-class feature, not a future roadmap item. They will publish threat models that cover issuer compromise, wallet loss, correlation attacks, and coercive verification. And they will make money from verification workflows, enterprise integration, and compliance automation rather than speculative identity tokens.
- Avoid permanent identity badges: non-transferable tokens are acceptable for low-risk reputation, but dangerous for regulated or personal attributes.
- Prioritize revocation latency: a credential system without fast, private revocation is unsuitable for finance.
- Separate wallets by context: employment, DeFi, gaming, and social credentials should not be forced into one correlatable address.
- Demand issuer governance: the registry matters more than the chain if nobody can remove a malicious signer.
- Use ZK selectively: prove predicates where privacy matters; do not wrap every policy in expensive circuits for branding.
Investors should be skeptical of projects selling identity as a token network before proving issuer adoption. The defensible value accrues to credential rails with real counterparties: exchanges, universities, payroll providers, government wallets, RWA issuers, and large consumer platforms. A beautiful DID method with no issuers is infrastructure for a ghost town.
Conclusion: The Future Is Private Credentials, Public Commitments
Decentralized identity will not replace passports, banks, or government registries. It will rewire how their attestations move through digital markets. The winning model is hybrid: institutions issue credentials, users hold and present them, verifiers receive only the minimum proof, and blockchains coordinate public commitments that make the system auditable without making people transparent.
The contrarian bet is that self-sovereign identity succeeds by being less on-chain, not more. Public chains should provide neutrality, censorship resistance, and verifiable state. They should not become permanent identity warehouses. If the industry gets this right, on-chain credentials can unlock compliant DeFi, portable reputation, bot-resistant communities, and privacy-preserving access control. If it gets it wrong, decentralized identity becomes the most efficient surveillance database ever built. The architecture choice is the market choice.