Confidential Transfers on Solana: Token Extensions' Privacy Layer Explained
Solana's Token Extensions include confidential transfers using zero-knowledge proofs. How they work, which tokens use them, and privacy limitations.
Solana's Token-2022 program introduced a privacy feature that most blockchains lack at the protocol level: confidential transfers. Instead of broadcasting transfer amounts publicly on-chain, the extension encrypts them using Twisted ElGamal encryption and proves their validity with zero-knowledge proofs. Validators confirm that balances add up and amounts are non-negative without ever seeing the actual numbers.
The result is a native privacy layer embedded directly in Solana's token standard: no mixers, no separate shielded pools, no bridge to a privacy chain. But the design comes with significant tradeoffs in what it hides, what it costs, and what it leaves exposed.
How Confidential Transfers Work
The confidential transfer extension is part of Solana's Token Extensions framework (the Token-2022 program). When a mint creator enables the extension at token creation, every token account for that mint can opt into encrypted balances. Each token account generates an ElGamal keypair that encrypts its balance and all incoming and outgoing transfer amounts.
Twisted ElGamal Encryption
Standard ElGamal encryption works on elliptic curves but produces ciphertexts that are expensive to operate on homomorphically. Solana uses a "twisted" variant that splits each ciphertext into two components: a Pedersen commitment (which supports efficient homomorphic addition) and a decryption handle (which lets the recipient recover the plaintext). This split is what makes it possible to verify that encrypted balances update correctly without decrypting them.
Transfer amounts are capped at 48 bits. Because ElGamal decryption requires a brute-force discrete log computation, larger values become impractical to decrypt. The protocol splits each amount into a 16-bit low part and a 32-bit high part, keeping decryption fast for both sender and receiver.
The Three Zero-Knowledge Proofs
Every confidential transfer requires the sender to generate three cryptographic proofs on the client side before submitting the transaction:
- Equality proof: demonstrates that the sender's ciphertext and the receiver's ciphertext encode the same transfer amount, despite being encrypted under different public keys.
- Ciphertext validity proof: confirms that the ciphertexts were properly constructed under the correct public keys and are not malformed.
- Range proof: a batched Bulletproofs-style proof certifying that both the transfer amount and the sender's remaining balance are non-negative. This prevents underflow attacks where a sender could create tokens out of thin air by transferring more than they hold.
These proofs are verified by Solana's native ZK ElGamal proof program. Because the combined proof data exceeds what fits in a single Solana transaction, the reference implementation spreads them across three transactions. Each proof is recorded in a temporary "context state" account, and the final transfer instruction references all three.
Compute cost matters: A 128-bit batched range proof alone consumes approximately 200,000 compute units. A standard SPL Token transfer uses roughly 2,000 to 4,000 compute units. Confidential transfers are roughly two orders of magnitude more expensive in compute, and they require multiple transactions to complete.
What Gets Hidden, What Does Not
The name "confidential transfers" is precise but narrow. The extension hides transfer amounts and account balances. Everything else remains public:
- Sender and receiver addresses are visible on-chain.
- The token mint is visible (observers know which token was transferred).
- Transaction timestamps and ordering are visible.
- The fact that a transfer occurred is visible.
This is a fundamentally different privacy model from systems like Monero or Zcash, which aim to hide sender, receiver, and amount simultaneously. Solana's approach is closer to Pedersen commitment-based confidential transactions originally proposed for Bitcoin by Gregory Maxwell, where the transaction graph remains visible but the values flowing through it are hidden.
| Property | Standard Transfer | Confidential Transfer |
|---|---|---|
| Sender address | Visible | Visible |
| Receiver address | Visible | Visible |
| Transfer amount | Visible | Encrypted |
| Account balance | Visible | Encrypted |
| Token mint | Visible | Visible |
| Transaction graph | Visible | Visible |
| Compute cost | ~3,000 CU | ~300,000+ CU (across multiple txns) |
The Auditor Key: Compliance by Design
One of the most distinctive features of Solana's confidential transfer extension is the optional auditor key. When creating a mint with confidential transfers enabled, the issuer can designate an auditor public key. Any transfer of that token then encrypts the amount not only to the sender and receiver but also to the auditor. The public sees nothing; the auditor sees everything.
This design directly addresses the regulatory tension around on-chain privacy. Rather than building privacy that is absolute (and therefore potentially problematic for KYC/AML compliance), the extension allows issuers to maintain a selective disclosure channel. A regulated stablecoin issuer, for example, could comply with travel rule requirements by granting regulators access to the auditor key while keeping amounts invisible to the general public.
The auditor key supports rotation, but rotation is forward-looking only. Amounts encrypted under a previous auditor key require retaining that old key to decrypt. If the auditor key is cleared entirely, only the individual account owners can see their own balances. The issuer guide recommends custodying the auditor secret key with the same rigor as a signing key.
Which Tokens Use Confidential Transfers
Adoption has been slow. The extension must be enabled at mint creation and cannot be added retroactively to existing tokens. This means major tokens minted under the original SPL Token program (including most USDC and USDT on Solana) cannot use confidential transfers without migrating to entirely new mints.
The most notable token with the extension configured is PYUSD, PayPal's stablecoin issued by Paxos on Solana. The PYUSD Token-2022 mint includes both the confidential transfer and transfer hook extensions. PayPal announced the confidential transfer capability in 2025, framing it as privacy that maintains regulatory transparency through the auditor key. However, whether the feature is actively switched on for end users remains unclear from public documentation: the mint has certain extensions initialized but not all are activated.
Ecosystem support is thin. As of late 2026, most Solana wallets display confidential balances as unavailable or zero, and most DEXs do not support trading confidential-transfer tokens. The lack of wallet and DEX integration creates a chicken-and-egg problem: issuers hesitate to activate the feature when users cannot interact with it seamlessly.
The 2025 Security Pause
In June 2025, security researchers identified a critical vulnerability in the ZK ElGamal proof system's Fiat-Shamir transformation. The flaw was severe: an attacker could have forged proofs to mint unlimited tokens or drain a victim's confidential balance.
The Solana Foundation responded by disabling confidential transfers on June 11, 2025, through a Token-2022 program update. Eight days later, the ZK ElGamal proof program was fully disabled via feature activation at epoch 805. A post-mortem published June 25, 2025 detailed the root cause and remediation steps. The fix required a full audit by Least Authority (completed November 2025) and a validator upgrade to Agave v2.1.21.
Confidential transfers returned to mainnet approximately one year later, in June 2026. The incident underscored a fundamental risk of on-chain privacy features: the cryptographic machinery is complex, and bugs in zero-knowledge proof systems can be catastrophic because they may be silently exploitable. With standard transfers, unauthorized minting is immediately visible in supply data. With confidential transfers, an attacker could inflate supply without detection.
Invisible exploits are the core risk: When transfer amounts are encrypted, supply inflation from a proof-system bug may not be detectable until someone tries to redeem. This is why the auditor key and regular attestations become critical for stablecoin issuers using confidential transfers.
How Solana Compares to Other Privacy Approaches
Every blockchain handles privacy differently. The tradeoffs involve what is hidden, how much it costs, and what trust assumptions are required.
Bitcoin: Transaction Graph Privacy
Bitcoin's privacy model focuses on the transaction graph rather than amounts. CoinJoin protocols mix inputs from multiple users to break the link between sender and receiver, but the amounts in each output remain visible. The anonymity set quality depends on participation, and CoinJoin transactions are 2 to 10 times larger than standard payments in virtual bytes.
Silent payments (BIP 352) address a different problem: they let receivers publish a single static address that senders derive unique one-time addresses from, preventing address reuse without out-of-band communication. Silent payments protect receiver privacy but leave amounts and spending patterns visible. For a deeper look at Bitcoin's evolving privacy toolkit, see our Bitcoin privacy landscape overview.
Ethereum: Shielded Pools and Privacy Rollups
Ethereum's privacy ecosystem is application-layer rather than protocol-native. Railgun operates as a smart contract privacy layer that shields ERC-20 token transfers using zero-knowledge proofs within a shielded pool. Users deposit tokens into the pool, transact privately within it, and withdraw to public addresses. The tradeoff is liquidity: privacy depends on pool size and usage volume.
Aztec takes a more ambitious approach as a dedicated ZK rollup with programmable private smart contracts. It uses a hybrid model with both public state (like standard EVM) and private state (encrypted via UTXO-style notes). Aztec's alpha mainnet launched in early 2026, aiming to bring full-stack privacy to Ethereum's execution layer.
Cross-Chain Privacy Comparison
| Feature | Solana Confidential Transfers | Bitcoin CoinJoin | Ethereum Railgun | Aztec |
|---|---|---|---|---|
| What is hidden | Amounts only | Sender-receiver link | Sender, receiver, amount | Sender, receiver, amount, logic |
| Addresses visible | Yes | Partially (mixed outputs) | No (inside shielded pool) | No (private state) |
| Protocol layer | Native token extension | Application-layer protocol | Smart contract layer | Dedicated L2 |
| Compliance hook | Auditor key (issuer-set) | None | Privacy Pools (selective disclosure) | Selective disclosure tools |
| Cost overhead | ~100x standard transfer | 2 to 10x standard transaction | Gas for shield/unshield + pool interaction | Rollup fees + proof generation |
| Trust assumption | Solana validators, proof system | Coordinator (some implementations) | Ethereum L1 security | Ethereum L1 + rollup sequencer |
Limitations and Open Questions
Graph Analysis Remains Effective
Because sender and receiver addresses are fully visible, blockchain analytics firms can still perform transaction graph analysis on confidential transfers. They cannot see how much moved, but they can see who transacted with whom and when. For many surveillance use cases (tracking sanctioned addresses, mapping business relationships, flagging suspicious patterns), the graph is more valuable than the amounts.
Retroactive Adoption Is Impossible
The extension must be enabled when the mint is created. Existing tokens on the original SPL Token program, including most of Solana's stablecoin supply, cannot adopt confidential transfers without creating new mints and migrating liquidity. This is a significant barrier when the goal is widespread privacy for tokens already in circulation.
Multi-Transaction Complexity
Requiring three transactions for a single transfer introduces UX and reliability challenges. If any of the proof-verification transactions fail (due to network congestion, insufficient compute budget, or transient errors), the transfer does not complete. Wallet developers must handle partial completion states, retry logic, and context-state account cleanup. The official integration guide provides helper functions that abstract this complexity, but the underlying multi-step flow remains a source of potential failure.
Client-Side Proof Generation
All three proofs must be generated on the sender's device. For desktop and server environments this is manageable, but mobile devices face computational constraints. Generating a batched range proof involves significant elliptic curve operations that can take several seconds on mobile hardware, impacting the perceived speed of what Solana markets as a high-throughput chain.
An Alternative Privacy Model: Off-Chain Transfers
Solana's confidential transfers hide amounts but leave the transaction graph fully exposed. A fundamentally different approach is to move transfers off-chain entirely, so that neither amounts nor the transfer event itself appears on the base layer.
This is the model used by Spark, a Bitcoin Layer 2 built on statechains. Spark transfers update key ownership off-chain without broadcasting any transaction. There is no on-chain record of who sent what to whom, no encrypted amounts to analyze, and no transaction graph to map. The privacy comes not from cryptographic concealment but from absence: what never touches the blockchain cannot be surveilled on it.
The tradeoff is different. Spark requires trust in its operator set (at least one of n operators must be honest during each transfer), whereas Solana's confidential transfers rely on the correctness of the ZK proof system and Solana's validator set. Each model suits different threat models and compliance requirements.
What Comes Next for Solana Privacy
Solana's privacy roadmap indicates that address hiding is in development, which would address the biggest gap in the current system. If both amounts and addresses can be concealed, Solana's native privacy would approach the level offered by dedicated privacy protocols.
Parallel efforts like Helius Privacy Rings (a separate system built on Token-2022 tokens using its own proof construction) and Arcium's multi-party computation approach are expanding the privacy design space on Solana. The ecosystem is moving toward multiple coexisting privacy layers rather than a single canonical solution.
For builders exploring privacy-preserving payment infrastructure, Spark's developer documentation covers how off-chain statechain transfers work at the protocol level, and the Solana vs Bitcoin L2 tradeoffs analysis provides a broader comparison of these two scaling philosophies.
This article is for educational purposes only. It does not constitute financial or investment advice. Bitcoin and Layer 2 protocols involve technical and financial risk. Always do your own research and understand the tradeoffs before using any protocol.

