The Post-Quantum Divide: Upgradeable or Trapped?

We may be asking the wrong question about quantum readiness.

The usual question is: “Which systems are quantum resistant?”

A more important question may be: “Which systems can become quantum resistant without breaking everything around them?”

That distinction is becoming visible in the cryptocurrency world.

Starknet is attracting attention because its programmable-account model allows the cryptographic rules protecting an account to change while the account itself can remain intact. Experimental work on Starknet’s Sepolia testnet has already demonstrated existing accounts moving to hybrid ECDSA + Falcon-512 authorization while retaining the same address and funds. Starknet itself says post-quantum transactions have already been demonstrated on mainnet, although broader network dependencies remain unresolved.

That is why Starknet’s recent governance discussion matters.

On October 22, a non-binding community poll closed with 93.52% of voting power supporting research into an independent post-quantum Layer 1. The vote does not approve such a transition. It gives StarkWare and the Starknet Foundation a mandate to study the technical and economic implications.

The underlying motivation is revealing. Starknet can change much of its own cryptography. But as an Ethereum Layer 2, it still relies on Ethereum for settlement, bridging, and data availability.

In other words: Starknet may be able to migrate faster than some of the infrastructure it depends upon.

That is not merely a blockchain problem. It is exactly the problem enterprises are about to encounter.

The first signal is the emergence of architectural crypto-agility.

Starknet’s account abstraction model separates the account from one immutable signature scheme. That means cryptographic controls can potentially be upgraded without replacing the account, moving assets, or forcing a network-wide hard fork. Starknet’s own recent analysis argues that this ability to change an existing account’s signature scheme in place is a significant advantage when planning for post-quantum migration.

This is still experimental territory. The Falcon migration demonstration on Sepolia is unaudited, and important issues remain around bridges, guardians, hardware signers, multisig, recovery, and protocol-level dependencies.

But those limitations actually make the experiment more useful. They demonstrate that post-quantum readiness is not achieved when one algorithm changes. It is achieved when the entire trust architecture can change with it.

The second signal is NIST.

NIST finalized IR 8547, Transition to Post-Quantum Cryptography Standards, on October 27. The report establishes the federal transition framework from quantum-vulnerable public-key algorithms toward post-quantum digital signatures and key-establishment mechanisms. It explicitly connects the transition to federal migration requirements and emphasizes that the process must extend through products, services, infrastructure, and the wider cryptographic supply chain.

The standards problem is no longer hypothetical. Organizations now have approved algorithms and increasingly defined transition expectations. The harder problem is implementation.

The third signal is the continued movement in quantum resource estimates.

Alice & Bob published a theoretical estimate indicating that its cat-qubit architecture could factor RSA-2048 with 22,593 physical qubits in approximately 3.14 days, while its ECDLP-256 estimate is approximately 29,588 physical qubits in 61.8 minutes under the paper’s assumptions. These are resource estimates, not demonstrated attacks, and the assumptions around error rates and fault tolerance are substantial. No machine capable of performing these attacks exists today.

But the important lesson is not the precise qubit count. It is that attack estimates are variables.

• Architectures improve.
• Error correction improves.
• Algorithms improve.
• Engineering assumptions change.
• The migration difficulty inside an organization, however, does not automatically shrink with them.

There is another signal worth watching. A Nasdaq-listed company, Next Technology Holding, recently authorized the possible sale or borrowing against approximately 5,833 BTC while citing exchange vulnerabilities and quantum-computing risk among its concerns. No sale is guaranteed, and quantum risk was not the only issue cited. But this is noteworthy because quantum exposure has now entered an actual corporate capital-allocation discussion rather than remaining confined to cryptography forums.

That is another step in the maturation of the issue.

The post-quantum transition may eventually divide organizations into two categories.

Not: quantum safe and quantum unsafe.  But:  upgradeable and trapped.

An upgradeable system can replace cryptographic mechanisms while preserving the business system around them. A trapped system cannot.

If changing a signature algorithm means replacing customer identities, rebuilding applications, changing hardware, migrating accounts, renegotiating vendor relationships, replacing certificates, modifying APIs, or redesigning workflows, the cryptographic problem has become an architectural problem. And that architecture determines migration time.

Consider what Starknet is experimenting with. An account can potentially retain:

• its address,
• its assets,
• its history,
• and its identity,
• while changing the cryptographic mechanism that authorizes transactions.

That is much closer to what enterprise crypto-agility should ultimately mean. Imagine the equivalent inside a bank:

• A machine identity survives while its signature algorithm changes.
• A customer relationship survives while its authentication mechanism changes.
• A payment system survives while its cryptographic controls change.
• A software distribution process survives while its code-signing mechanism changes.
• A device survives while its certificates and trust roots change.
• The business object remains.
• The cryptography underneath it evolves.

That is a fundamentally different capability from simply having ML-KEM or ML-DSA available somewhere in the technology stack. It also explains why NIST’s emphasis on visibility, supply chains, migration planning, and crypto-agility is so important.

The question is no longer: “Do we have access to post-quantum algorithms?”

We do.

The question is:  “Can we introduce them without breaking the systems that depend on the cryptography they replace?”

That is a much harder test. And it gives boards a surprisingly simple question to ask management: “If we had to replace one of our critical cryptographic mechanisms tomorrow, what else would we have to replace with it?”

If the answer includes customer accounts, major applications, devices, vendors, or core infrastructure, the organization has discovered something important. It may have strong cryptography. But it does not yet have crypto-agility.

That distinction will matter more as the two clocks continue moving. The quantum capability clock. And the migration clock. The organizations best positioned for the transition will not necessarily be those that correctly predict Q-Day.

They will be those whose architecture allows them to respond when the prediction changes.

Think Earlier. Trust Must Survive.