Two edges of the same wave

Quantum resilience and AI discipline are usually written as separate problems. They aren’t. 

For years, post-quantum cryptography lived in research papers and central-bank working groups. Most companies assumed they had time. Most assumed wrongly. 

The shift has a name. Harvest now, decrypt later. Encrypted data is being collected and stored today – not because anyone can decrypt it yet, but because the calculation is that one day someone will. Once the cryptographic ceiling lifts, the archive opens. For systems that depend on long-lived trust – financial, medical, identity, sovereign – that is not a problem you can patch your way out of in retrospect. 

What is actually new – and who is saying it 

Most quantum-resilience writing comes from one camp: cryptographers warning the wave is coming, or quantum researchers explaining why it is not here yet. Both are correct in narrow ways. Neither view is the one you act on. 

We work at both edges of the same shift. Through initiatives like Unveil, we sit close to teams building the admissibility layer that decides what should be computed before any work begins – the discipline of teaching AI to say no when the answer does not exist. Through our partnership with QSentinel, we deploy quantum-resistant cryptography into the workspace layer where institutional and enterprise data actually lives. 

Two different problems. One operator vantage. Close enough to know what is coming, and in production with the parts that have to change. 

The calendar is not the timeline 

Press releases about quantum hardware tend to talk in decades. The decisions in front of you do not run on that clock. 

If your data needs to remain confidential for ten years – financial records, healthcare files, defence procurement, IP, anything that ages into liability – then the relevant date isn’t when a fully fault-tolerant quantum computer ships. It’s when the data you are encrypting today loses the protection it was sold with. That moment is closer than the calendar suggests, and the harvest-now-decrypt-later strategy is the proof. The question is not whether quantum systems will eventually matter to you. It is whether the data you are generating right now is already on a timeline you do not control. 

This is an infrastructure problem, not an algorithm problem 

The temptation, when a security problem looks technical, is to think the answer is technical too. Pick a post-quantum algorithm, swap it in and move on. 

That is not how this resolves. Cryptography lives everywhere – in payment rails, in identity systems, in machine-to-machine authentication, in the boring infrastructure no one looks at until it breaks. Replacing it is a coordination problem across teams, vendors, regulators, and decades-old systems that were never designed to be cryptographically agile. The hard part is not choosing the algorithm. It is knowing where it lives, what it touches, and how to migrate it without breaking the business running on top of it. 

The AI side – saying no when the answer isn’t there 

The same era is reshaping AI in a related, often missed way. 

Most AI infrastructure is built to say yes. More data. More models. More inference. The default answer to any prompt is try harder. But more is not always progress, and every yes has a cost – capital, infrastructure, energy, time. Some of the most expensive compute in the world is spent on candidates that were never physically possible to begin with. 

Initiatives like Unveil are pushing on this. The idea is straightforward: a gate that sits in front of computation and decides whether a viable answer can exist before the work begins. Approve. Transform. Reject. The discipline is not speed – it is intent. Compute less but compute only what is worth computing. For founders watching their cloud bills compound, and for institutions running large AI estates, this concern has arrived much sooner than expected. 

The quantum side – the layer most people overlook 

 

Quantum-resistant cryptography is real, standardizing fast, and increasingly available. The challenge is rarely the cryptography itself. It Is getting it into the layer where the data lives. 

Through our partnership with QSentinel – a Swiss cybersecurity venture building post-quantum cryptography for the enterprise workspace – we deploy and integrate this layer into the systems most companies forget about: documents, email, chat, AI workflows, sovereign communications. The unglamorous infrastructure that doesn’t show up on a perimeter diagram but holds most of an organization’s actual intelligence. 

This is where resilience either lives or does not. A patched perimeter is still vulnerable if the workspace layer behind it is not migrated. Deployment – not invention – is where the threat is actually neutralized or not. That is the work. 

What this means, depending on who you are 

If you are an institution, the question is partnership selection. Pick partners who sit on both sides of the shift – close to where the frontier is being defined, in production with the parts that have to change. Readiness assessments and white-paper roadmaps fall short of the actual work. What matters: protecting your workspace layer and maintaining sustainable compute discipline. 

If you are a founder or operator, the discipline is not different – it is earlier. The architectural choices you make now decide whether your business is resilient by design or retrofitted later under pressure. The strongest systems carried resilience in their original design, long before rising costs made those decisions harder. 

Early, not exposed 

 

The quantum era won’t arrive as a single moment. It will arrive as a slow shift in what data stays safe, what compute is intentional, and which infrastructure outlasts the wave. 

The organizations that come through it cleanly won’t be the ones who saw the future first. They will be the ones who built around two disciplines – saying no to wasteful compute and protecting the parts of their stack they would otherwise leave exposed. 

We help on both sides. Early, not exposed.