Inteltech
The Quantum Risk Is Your Data's Shelf Life
Whether quantum computing arrives in five years or twenty, data captured today is at risk now. The migration clock starts at your retention period.
The quantum question is usually asked as a date: when will a quantum computer be able to break RSA? Nobody knows, the estimates move in both directions, and waiting for an answer is a way of not making a decision.
The better question is a duration: how long does the data you are protecting have to stay secret?
Those two numbers, the arrival date and the shelf life, are the entire risk calculation. If the shelf life is shorter than the arrival date, you have time. If it is longer, you are already exposed, because the attack does not begin when the machine is built. It begins when the traffic is recorded.
The attack that has already started
Encrypted traffic can be captured and stored cheaply. A patient adversary records what it cannot read today and waits. When a cryptographically relevant quantum computer exists, the archive becomes readable in retrospect.
That is the harvest-now-decrypt-later model, and it is the reason the standard advice sounds urgent even though no quantum computer can break RSA today.
The mathematics is not in dispute. Shor's algorithm, published in 1994, would break RSA and elliptic-curve cryptography on a sufficiently large quantum computer. The uncertainty is engineering, not theory: how many physical qubits, how much error correction, how long.
Symmetric encryption is a different case. A quantum computer running Grover's algorithm reduces the effective strength of a symmetric key but does not break it, which is why the guidance for AES is to use longer keys rather than to replace the algorithm. The migration problem is concentrated in public key cryptography: key exchange, signatures, and the certificate infrastructure built on both.
Why the Y2K comparison fails
The comparison to Y2K is tempting because both are about a deadline and a lot of old systems. It is also misleading in a way that leads to bad planning.
Y2K had a fixed date, a known failure mode, and a fix that could be tested. The work was large and the scope was knowable. Post-quantum migration has no fixed date, a standard that is still being adopted, and a fix that depends on hardware and software you do not control.
That changes the planning method. You cannot run a remediation program against a date. You can run one against an inventory, a priority order, and a capability to change algorithms when the standards settle. The organizations that treat this as a Y2K project will spend the money early and discover they have to do it again when the guidance moves. The organizations that build the capability will absorb the change.
Shelf life sets the priority
Sort your data by how long it has to stay confidential, and the priority order writes itself.
| Data | Shelf life | What it implies |
|---|---|---|
| Session tokens, live traffic keys | Minutes to hours | No harvest risk worth modeling |
| Transaction records | Years | Migrate on the normal refresh cycle |
| Personnel and health records | Decades | Plan migration now; archives included |
| Design data, source code, legal holds | Decades | Highest priority; often forgotten |
| Anything with a statutory retention | Set by statute | The retention period is the exposure window |
The rows that surprise people are the last two. Design data and source code are intellectual property with a long competitive life, and they are frequently stored in systems nobody classified as sensitive. Legal holds and archives are worse, because they are copies made deliberately to be kept, often with the encryption of the day and no plan to re-encrypt.
Backups deserve their own line. A backup encrypted with a legacy algorithm is a data set with the original exposure plus however long you keep the tape. Retention schedules written before this problem existed are a good place to start the review.
Find where the cryptography actually is
The inventory is the deliverable that makes everything else possible. It is also larger than most teams expect.
Start with the obvious: TLS endpoints, VPN concentrators, certificate authorities, and the internal services that terminate their own connections. Then keep going, because public key cryptography is also in code signing, package repositories, firmware images, smart cards, database encryption, backup appliances, and the identity provider that everything else trusts.
For each entry, record four things: the algorithm, where the key lives, who can change it, and what depends on it. The fourth column is the one that turns an inventory into a plan, because a certificate authority with a hundred dependent services is a different project from a single appliance.
The uncomfortable finding is usually not the number of systems. It is how many of them cannot change an algorithm without a vendor releasing firmware, and how many vendors have not answered the question.
What ready looks like
Readiness is not a completed migration. It is a set of properties.
Cryptographic agility: the ability to change an algorithm without rewriting the application. Hybrid mode: the ability to run a classical and a post-quantum algorithm together, so that breaking one does not break the connection, and so that you can interoperate with a counterparty that has moved at a different speed. Validation: a path to a validated cryptographic module, because a correct implementation that is not validated is not usable in a regulated environment.
None of those are purchases. They are design decisions, and they are cheaper to make before the hardware refresh than after.
Executives ask what this costs. The honest answer is that the cost is mostly the work you were going to do anyway, done in a different order.
You will refresh the certificate authority. You will replace the VPN concentrators. You will re-encrypt the archive when the storage platform changes. The post-quantum question is whether those events are sequenced against a shelf-life calculation or against a vendor's release schedule.
Start with the data that must stay secret the longest. Put the inventory next to it. Then decide what to change first, and write down the decision, because the next person to ask will want to know why.
The quantum computer is not the deadline. The deadline is the retention period on the data you are protecting today.