Post-quantum is a resilience problem – treat it like one
I spent years writing threat assessments for senior people who needed to know what a new capability meant. The first question was always what a capable adversary would do with it.
For a quantum computer able to break today’s encryption, the familiar answer is reading secrets they’ve already collected. The answer that worries me more is that they could forge the signatures your systems rely on to know an instruction is genuine.
When you can’t tell real from fake
Signatures tell a payment system an instruction is genuine. They tell a protection relay that a firmware update came from its manufacturer, and a smart meter that an order to cut supply came from the energy supplier. Once they can be forged, an attacker doesn’t need to knock anything offline. Operators may have to do it themselves, because they can no longer tell real from fake.
Banks, energy networks and telecoms operators all answer to a resilience regime now. Few of the scenarios I’ve seen test for the day signed instructions stop being trustworthy. Ask whether yours does.
Why now
Every estimate of the machine needed to break RSA-2048 has moved the same way. In 2019 it was around 20 million qubits. In 2025 it was around a million. Nobody can put a date on it.
The defenders’ timetable is already written. NIST expects this migration to take more than a decade. The NCSC wants discovery done and a plan in place by 2028, and that’s two budget cycles away. From January 2027, new US national security systems must be quantum-resistant. Supplier roadmaps are being set now, around the customers who asked first.
Then there’s asset life. A network switch or substation controller installed this year will still be running in the 2040s. If it can’t change algorithms, you’ve already bought its replacement.
Most of the fix is in someone else’s hands
Cryptography sits in platforms, hardware security modules, cloud services, operational technology and supplier products. You carry the risk and control a fraction of the fix. Every month spent waiting for a regulator to set a date is a month less to influence your suppliers.
The first change won’t be the last
In July, an AI model helped a researcher find a weakness in HAWK, a post-quantum signature scheme in the third round of NIST’s process. It took around 60 hours, and the designers withdrew HAWK within a day. It was never deployed, and the finalised standards are unaffected. But cryptography is now being tested faster than ever, and this migration will be the first of several. The capability worth funding is the ability to change algorithms without breaking the business.
Money is the blocker
Our roundtables with banks and payment firms keep ending in the same place. Nobody doubts the maths. Getting it funded is the hard part, and it’s harder while quantum competes with AI for board attention. We helped two UK energy network operators build a business case phased against the NCSC milestones, and their migration is now funded within their regulatory settlements.
Where I’d start
Give one executive ownership, ideally whoever already answers for resilience.
Write crypto-agility into procurement now. It costs almost nothing and stops you buying more of the problem.
Map your exposure, starting with what is hardest to change and what must stay trusted longest.
Add quantum scope to programmes that already have funding.
Start now and you set the pace and cost of your own transition. Leave it, and suppliers, regulators and eventually an adversary will set it for you.
