Can Solana Go Quantum-Safe Without Sacrificing Its Speed?

SOL
Post-Quantum CryptographyQuantum ComputingFalcon signaturesaccount migrationQuantumglowAlpenglowconsensus
6 hours agoSource: crypto.news
Can Solana Go Quantum-Safe Without Sacrificing Its Speed?

Solana’s developers are preparing for quantum computers by testing new signatures, redesigning how validators prove agreement and developing a way to move existing accounts. Preserving fast settlement is part of that work, but a research design that keeps consensus quick still has to survive the demands of wallets, applications and the live network.

Summary

  • Solana activated Alpenglow in October, while the proposed Quantumglow adaptation remains separate research rather than a deployed post-quantum security upgrade.
  • Current Ed25519 transaction signatures occupy 64 bytes, compared with 666 bytes for Falcon-512 in Jump Crypto’s published comparison.
  • Solana’s documentation lists a 1,232-byte transaction limit and a 4,096-byte limit for the v1 format, making format assumptions important.
  • Quantumglow would replace compact aggregate certificates with locally recorded evidence and an approval mechanism designed to preserve fast finality.
  • Anza’s account-migration prototype uses proof of existing seed material, addressing a different problem from replacing the signatures used by validators.

Solana has spent years making transactions faster, cheaper and easier to process at scale. Preparing for quantum computing now requires its developers to revisit some of the cryptography that made those performance choices possible.

The problem is not confined to selecting a replacement for the signature attached to a payment. Validators use cryptography to communicate, authenticate blocks and agree on the ledger, while wallets and applications use it to establish who can move assets.

Developers have proposed different solutions for those jobs, including Falcon signatures, an account-migration proof and a post-quantum adaptation of Alpenglow called Quantumglow. Their progress gives Solana a preparation path, without establishing that the network has already completed that transition.

The distinction has become easier to miss since Alpenglow went live in October. Solana has deployed a faster consensus system, while the research intended to protect that system from a sufficiently powerful quantum adversary remains a separate undertaking.

Alpenglow’s launch did not complete the security transition

Solana’s Alpenglow upgrade documentation records activation in epoch 1064 on Oct. 21. The Foundation describes a target of approximately 150 milliseconds for finality, compared with roughly 12.8 seconds under the previous TowerBFT system.

Finality measures when agreement becomes irreversible under the protocol’s assumptions. It should not be read as a promise that every wallet action, exchange deposit or application workflow completes within 150 milliseconds.

The October migration plan concerned compatible validator software and operational preparations for changing consensus. It did not announce a replacement for all the cryptography securing user accounts.

The upgrade has a defined scope. Alpenglow changes how validators agree, while leaving transaction execution, programs and fee mechanics in place.

Its rollout is divided again between Votor, the voting component now live, and Rotor, a planned data-distribution component. Describing every part of the research roadmap as activated would combine work at different stages.

Quantum preparation introduces another distinction because a faster voting system can still rely on signatures vulnerable to a future quantum attack. Speed and the mathematical assumptions behind authorization are different properties.

A successful migration must preserve both the agreement between honest validators and the practical ability of ordinary users to authorize transactions. Protecting one while leaving the other exposed would leave assets vulnerable through the remaining route.

For an exchange, the difference reaches beyond whether a deposit receives a quick confirmation. Its custody software must generate accepted signatures, its recovery arrangements must remain usable and its monitoring systems must recognize the new transaction behavior.

For an application, a consensus upgrade can be largely invisible while a change in supported signatures requires integration work. Programs that verify signatures directly may need their own changes even after the network accepts a new transaction format.

The deployed Alpenglow system therefore supplies a performance baseline for future testing. It does not supply evidence that Quantumglow, Falcon account migration and every application dependency have already been tested together under production conditions.

The threat is forged authorization, not a faster password guess

The relevant quantum risk concerns the public-key mathematics used to verify signatures. A sufficiently capable machine could use Shor’s algorithm against elliptic-curve systems, allowing an attacker to derive signing secrets from public information.

An attacker could then impersonate the legitimate signer. A transaction could carry a mathematically valid legacy signature even though the legitimate account owner never approved it.

The National Institute of Standards and Technology has published post-quantum signature standards intended to resist known attacks from both conventional and quantum computers. Systems still have to implement those algorithms through software, operating procedures and rules for handling assets authorized under older schemes.

Dustin Moody, who leads NIST’s post-quantum standardization work, put the preparation case plainly:

“We encourage system administrators to start integrating them into their systems immediately, because full integration will take time.”

NIST finalized ML-DSA in 2024 alongside other initial post-quantum standards. Key establishment and digital signatures serve different purposes, so adopting an encryption or key-exchange standard would not by itself replace a blockchain’s transaction authorization scheme.

The agency’s initial standards announcement urged administrators to begin integrating the tools because complete integration takes time. The agency’s message focused on starting the transition early enough to accommodate the work required across existing computer systems.

The same distinction appears in the European watchdog warnings issued this year. The authorities emphasized the time needed to migrate financial systems without saying that practical attacks on present blockchain cryptography were already available.

An attack on signatures differs from collecting encrypted communications now and decrypting them later. Blockchain transactions are generally public already; the central concern is whether someone could impersonate an owner or validator in the future.

Public exposure matters because a conventional signing system assumes that publishing a verification key does not reveal the secret needed to sign. A quantum attack would undermine that assumption for vulnerable algorithms.

Moving coins into a new application without addressing the network’s underlying authorization can leave another vulnerable route open. The security claim must follow the entire path through which a transaction becomes accepted.

There is no reliable date in the cited Solana preparations for a machine that can carry out such an attack. Their work addresses the lead time needed to replace deployed systems and maintain access for users whose assets depend on existing keys.

Signature size changes the economics of a transaction

Solana’s transaction format documentation lists a 64-byte Ed25519 signature for each signer. A transaction carries instructions, account references and other information, leaving a finite amount of space for everything needed to execute it.

Jump Crypto’s published migration-path comparison lists a 666-byte Falcon-512 signature and a 2,420-byte ML-DSA-44 signature. Those are different engineering choices even before their public keys and integration requirements enter the calculation.

Replacing a 64-byte signature with 666 bytes adds 602 bytes for one signer. For a transaction requiring three separate signatures, the raw difference would be 1,806 bytes before any other format changes.

Size alone cannot predict transaction fees or network slowdown. A larger signature can verify quickly, while a compact signature can require more computation or introduce implementation difficulties.

At a large transaction count, the raw difference becomes easier to see. One million authorizations carrying an additional 602 bytes would add 602 million bytes, approximately 602 decimal megabytes, before accounting for any other changes.

The calculation is deliberately limited to signature data. Actual network traffic depends on how many times information is relayed, whether keys or proofs are reused and how the protocol packages messages.

It helps explain why a design can benefit from smaller signatures even when a competing signature verifies quickly. Bandwidth must still be purchased and provisioned, with operators at the edge of their connection capacity facing a different constraint from a laboratory machine on a fast local network.

A benchmark publishing verification time alone would miss that cost. A benchmark publishing bytes alone would miss the computational cost, making both necessary to assess the promised performance.

The public key creates another expense. Jump’s comparison lists 897 bytes for a Falcon-512 public key, against 32 bytes for Ed25519, raising the question of where those keys are stored and when they must be transmitted.

An account format that stores a commitment to a larger key can reduce recurring address overhead, although verification still needs a secure way to obtain and check the relevant key. The design must specify where that information is held and how its authenticity is established when it is needed.

Transaction limits make the choice concrete. The Solana documentation gives 1,232 bytes for the existing packet-sized limit while identifying a 4,096-byte limit for the v1 format.

An ML-DSA-44 signature alone exceeds 1,232 bytes but fits within 4,096 bytes, before adding the rest of a transaction. That observation makes the applicable format essential to any claim that a signature can or cannot fit.

A bigger permitted transaction does not make bandwidth free. Validators must receive the extra data, users must transmit it and the network has to handle the resulting workload during busy periods.

A fair comparison needs the same transaction content on both sides. Testing an ordinary one-signature transfer against a complex trade with several signatures would combine the cryptographic cost with differences in the underlying operation.

Wallets can face costs that validator benchmarks miss. A signature scheme may be fast to verify on a server but more demanding to generate safely on a phone, browser extension or hardware signing device.

Falcon offers a smaller signature, with implementation work attached

Anza and Firedancer introduced early Falcon implementations in April. The initial Falcon testing gave the client teams a working candidate to evaluate before any decision about migrating the live account system.

The Foundation’s quantum readiness statement said the teams had reached similar conclusions through independent work. Their stated expectation was that a migration could be managed without a meaningful performance impact when activation became appropriate.

The Foundation said in April that “network performance is not expected to see a meaningful impact.” Its claim expresses the developers’ expectation for the migration and leaves the complete workload to be assessed through implementation and testing. An implementation in a repository establishes progress, while a deployed system has to handle users, adverse traffic and the variety of devices that sign transactions.

Falcon’s attraction follows partly from its relative compactness among the candidates under consideration. A smaller signature leaves more room for transaction instructions and reduces the amount of new data attached to each authorization.

Jump’s discussion separates verification from signing. Verification can use integer-only arithmetic, while producing signatures safely requires careful implementation and attention to side-channel risks.

Signing off-chain keeps that work away from the blockchain’s execution budget, but someone still performs it. Wallet providers, institutional signing systems and hardware devices become responsible for implementing and auditing that step.

For a user, the practical test is whether approval remains reliable on their actual device. Repeated failures, unacceptable signing delays or recovery incompatibilities would affect usability even if validators verify the resulting signature quickly.

For a custodian, a new algorithm can require changes to approval policies, key-generation procedures and recovery testing. Systems built around a particular conventional signature scheme may not support a replacement through a simple configuration change.

Mixed support creates an additional transition period. A wallet might understand the new signature while an exchange, application or connected signing device does not yet accept it.

Migration planning must therefore identify the capabilities required before legacy signing can be restricted. Removing the old method too early could strand legitimate users; leaving it available after it becomes forgeable would preserve the attack route.

The existence of multiple candidate schemes gives researchers options as performance and security evidence changes. It makes explicit specification important, since the label “post-quantum” does not identify one universally interchangeable transaction format.

Quantumglow changes the evidence validators exchange

Wallet signatures are only part of the problem. Alpenglow uses compact aggregate signatures to represent agreement from multiple validators, a feature that helps it communicate decisions efficiently.

Anza’s Quantumglow research overview describes a different approach because an equally suitable post-quantum aggregate was unavailable under its requirements. Validators would record certificates locally after collecting enough evidence, with an approval mechanism supporting the communication between them.

The Quantumglow proposal coverage in October separated that work from the Alpenglow activation. Promoting an adaptation and explaining its research does not establish an approved deployment date.

Quentin Kniep, Kobi Sliwinski and Roger Wattenhofer published the overview in July, after the updated Alpenglow paper introduced its post-quantum section. Their proposed signature design, Ax, is tailored to validator communication, which has different requirements from an ordinary user’s authorization of a payment.

The construction uses hash-based signatures and advance exchange of one-time public-key material. The proposal consequently moves some work ahead of the message that has to be processed quickly, making preparation and cached information part of the performance story.

The approach requires more than enlarging an existing transaction’s signature field. A protocol can change its communication pattern to accommodate different cryptography, with new implementation obligations for the operators participating in it.

For a reviewer, the question becomes whether those preparations continue working when an operator restarts, communication is delayed or validators change. Fast processing under a stable setup is useful evidence, while reliable handling of changing conditions determines how that design behaves as a service.

The underlying engineering question is how much evidence each validator must receive to trust a decision. Sending a compact aggregate once and receiving many independent messages can prove related facts while imposing different communication patterns.

A design that removes a large certificate from the normal path can preserve speed without making every message smaller. It can change when evidence is assembled, who retains it and what must be communicated during an exceptional event.

The proposed speed therefore depends on the conditions under which validators communicate. The ordinary path under healthy participation may remain fast while fallback behavior during faulty or delayed communication takes longer.

The distinction is familiar in consensus research because speed during normal operation and the ability to keep making progress under disruption are separate requirements. A usable design must describe both, including the conditions that push validators from the normal path into a slower recovery process.

A network benchmark should consequently show the distribution of finality times. An attractive average could conceal a long tail of slower confirmations during conditions that matter most to traders and payment services.

Bandwidth measurements need more than the size of an individual message. A short message sent to many peers can create substantial total traffic, while caching or reuse can reduce later costs.

The authors distinguish fast and slower certificate events, preserving the normal fast path while providing the evidence needed for robustness. Their account describes the fast case in terms of votes representing 80% of stake, making the participation condition part of the claimed result.

A percentage of stake is different from a percentage of machines. Fifty small validators and one large validator can contribute different amounts of voting weight, so a test’s participant count alone does not describe the agreement being measured.

Disclosing the distribution of stake would therefore make fault tests more informative. Delaying a large participant can affect the fast path differently from delaying several operators whose combined weight is much smaller.

Independent implementations would help establish whether the design can be reproduced outside the authors’ setup. They would expose assumptions about networking, hardware and operator behavior that are difficult to evaluate from a simulation alone.

A formal security argument and an implementation benchmark answer different questions. The former examines whether the protocol satisfies stated properties under its assumptions; the latter measures what happens when software performs the work on specified machines.

Neither replaces the other. A fast implementation without adequate security reasoning would fail the purpose of migration, while a sound design with impractical resource requirements would be difficult for validators to operate competitively.

Existing accounts need a route that an attacker cannot copy

Changing the signature accepted for newly created accounts leaves existing holdings to deal with. A user who still controls an old wallet needs a way to establish ownership that remains convincing after the old signature scheme becomes vulnerable.

Anza’s quantum security assessment describes a prototype that proves knowledge of the seed material underlying an Ed25519 account and binds it to a new post-quantum public key. That migration proof serves a different function from authorizing an ordinary payment.

The approach exploits a distinction between recovering a derived signing secret and recovering the original seed. Under the proposed construction, an attacker breaking the conventional signature would not automatically possess the seed knowledge needed for migration.

The policy surrounding that proof remains important. If a vulnerable legacy signature continues to authorize unrestricted spending, an attacker may still use that route rather than attempt the migration proof.

The prototype therefore should not be read as an activated recovery guarantee for every account. Its deployment would require compatible rules, implementation and a decision about legacy authorization.

Anza reports an approximately 400-kilobyte proof for its initial prototype. At roughly 400,000 bytes, that is about 98 times a 4,096-byte transaction budget, illustrating why migration-proof handling is a separate engineering task.

The initial proof size leaves an implementation problem to solve. Proof compression, staged submission or changes to the verification process could alter how the work reaches the network.

Migration is a different workload from daily payments. An expensive operation performed once to protect an existing account can be acceptable where the same expense on every transfer would be impractical.

The users least prepared to act are an important part of the decision. Dormant accounts, outdated wallets and institutional approval systems can remain dependent on old authorization long after a new option becomes available.

Ledger CTO Charles Guillemet’s discussion of Bitcoin migration timelines highlights that coordination problem across crypto. Selecting cryptography can be faster than updating custody, wallets and the holdings they control.

For Solana, keeping familiar addresses could reduce some disruption, but applications would still need to recognize the new authorization rules. Preserving an address does not mean the software responsible for signing can remain unchanged.

Wallet vaults cannot protect every route into the ledger

Application-level vaults offer users a way to add post-quantum checks before a complete protocol transition. A program can require an additional proof or signature before releasing assets under its control.

Assets actually placed inside the vault can benefit from the additional check. The protection depends on what the program controls, how it is implemented and which network assumptions continue to hold underneath it.

A vault does not independently rewrite validator consensus. It cannot change every other application that accepts an elliptic-curve signature as evidence of authorization.

The fee payer creates a practical example. A transaction still needs to pay network fees through the account model accepted by the protocol, leaving that account relevant even when the main asset balance sits behind additional checks.

The source of the security claim consequently matters as much as the headline description. A protected application, a protected transaction-signing path and a protected consensus system describe different achievements.

An exchange considering a vault would need to examine its withdrawal workflow and operational limits. A user considering the same program would need wallet support and a recovery process that actually works with the additional authorization.

More complicated authorization can introduce ordinary software risks alongside the quantum problem it is intended to address. Contract bugs, lost secrets and approval mistakes remain possible after a stronger signature is added.

A protocol facility for verifying a new signature would help application developers avoid rebuilding the same cryptographic operation independently. It would not decide when a particular program should require that signature, which key it trusts or how an owner recovers access.

A trading program accepting instructions from an external service would still need to authenticate that service correctly. Replacing the algorithm inside one verification operation cannot correct an authorization policy that trusts the wrong party.

Wallet interfaces then have to describe approvals accurately enough for users to understand the action. Stronger mathematics does not prevent someone from signing a malicious instruction that the interface presents as harmless.

Testing should include those routine failures because a migration people cannot operate safely can create losses without any quantum computer being involved. Recovery behavior belongs in the security review before activation, with testing covering what happens when an owner’s normal signing process fails.

The same applies to bridges and programs using embedded signature checks. A protocol-level upgrade does not automatically replace cryptographic assumptions written into every existing application.

A credible network inventory would identify where vulnerable schemes are used and which team can change each dependency. It would make progress measurable without describing one component’s deployment as completion of the entire project.

The inventory can help prioritize work. Assets with high values, exposed authorization and difficult recovery procedures may warrant earlier integration effort than low-value experimental applications, subject to their actual threat model.

Speed claims need a complete workload and a deployment record

Solana’s preparation provides several avenues for preserving performance, but the evidence must be matched to the claim. Fast verification of one signature does not establish unchanged transaction throughput, and a consensus simulation does not establish wallet migration at scale.

The useful benchmark would disclose the client version, hardware, network conditions, signature scheme and transaction format. It would identify the mix of transfers, trades and other operations being processed.

Multi-signature workloads deserve their own results because the added data can scale with the number of authorizations. A system tested only with one-signature transfers may understate the effect on institutional or application workflows.

Operators would need to see CPU use, memory, bandwidth and storage alongside finality. An unchanged confirmation time achieved through much more expensive infrastructure would still change the economics of running a validator.

A larger resource requirement can affect who remains able to participate. If only operators with faster connections or more expensive hardware can maintain competitive performance, an apparent speed success may carry a participation cost.

Ordinary users experience a different set of measurements: signing time, transaction size, fee, failure rate and recovery compatibility. Publishing those alongside validator results would show whether performance has been preserved across the whole service.

The final comparison needs a clear baseline. Alpenglow finality, transaction execution capacity, block timing and exchange crediting are separate measures that should not be combined into one claim about speed.

Research can establish a plausible path before every implementation choice is settled. Deployment requires the additional decisions that tell operators and users exactly what changes, when it activates and how existing accounts continue to function.

As of the reporting reviewed for this feature, the public record separates deployed Alpenglow from proposed post-quantum adaptations. The work needed to close that gap includes client changes, account rules, application support and an activation process that the network can verify.

What to Watch

  • Activation proposals: Published specifications identifying which post-quantum changes are proposed, approved, implemented in clients and activated on the live network.
  • Benchmark conditions: Reproducible results covering finality, bandwidth, signing, verification and transaction failures under the same hardware and workload assumptions.
  • Account migration: The verification cost, submission method and recovery rules for binding existing accounts to new post-quantum authorization.
  • Wallet compatibility: Support across phones, hardware devices, exchanges and institutional signing systems before any restrictions on legacy signatures take effect.
  • Validator costs: Whether preserving fast finality requires additional infrastructure spending that changes the economics of independent network participation.

FAQs

Is Solana already quantum-safe?

The public preparations reviewed here do not establish a completed network-wide transition. Alpenglow is live, while Quantumglow, signature replacements and account migration involve separate work. A claim about one component should specify its implementation and activation status.

Would a quantum attacker need a wallet’s recovery phrase?

The signature threat concerns deriving signing secrets from public-key information using a sufficiently capable quantum computer. It does not require guessing a user’s phrase in the ordinary sense. Solana researchers are examining migration proofs that rely on underlying seed knowledge an attacker would not automatically recover through that attack.

Why does signature size matter for Solana?

Signatures consume transaction space and must travel across the network. Replacing a 64-byte signature with a larger one can increase bandwidth requirements and reduce the remaining room for instructions. The applicable transaction format, number of signers and verification costs all affect the result.

Does a larger signature always mean slower verification?

Size and verification time are different measurements. A scheme with larger signatures can verify efficiently while creating a bandwidth or storage cost elsewhere. Benchmarks must measure the complete workload rather than infer network speed from signature length alone.

What is Quantumglow intended to change?

Quantumglow addresses how validators authenticate and establish agreement in a post-quantum adaptation of Alpenglow. Its proposed communication changes avoid relying on an identical replacement for today’s compact aggregate certificates. The research does not by itself migrate user wallets or establish a mainnet activation date.

Why can migration use a more expensive proof than a payment?

Migrating an account can be an occasional operation, while payments may occur repeatedly. A larger one-time proof can therefore have a different economic role from a large signature attached to every transaction. Submission, verification and congestion costs still need workable implementation rules.

Can a quantum vault protect the entire network?

A vault can add authorization requirements for assets placed under its program’s control. It cannot independently replace consensus, networking or every vulnerable signature check in other applications. Fee-paying accounts and recovery procedures remain part of the security assessment.

What evidence would show that Solana preserved its speed?

Comparable benchmarks should use disclosed hardware, client versions, transaction formats and workloads, with results for users as well as validators. A confirmed deployment record is then needed to connect those tests to the live system. This is educational analysis, not investment advice.