Resilience by design, not by detection

A growing share of attacks cannot be patched in time. That changes what continuity in critical environments has to be built on.

By Rob Pike, Founder and CEO, Cyemptive Technologies

The threat that arrives before the advisory does

The most pressing threat today is not a malware family. It is the shrinking window between an attack being used and an attack being knowable. In the first half of 2026, 23% of all known exploited vulnerabilities were used on or before the day the CVE was published, according to VulnCheck’s State of Exploitation report. For that quarter of cases, no patching speed would have helped: there was nothing to install.

Meanwhile the volume climbs. FIRST forecasts 59,427 new vulnerabilities for 2026 — roughly one every nine minutes, and the first year above fifty thousand. You cannot patch what has not been published yet.

For critical infrastructure, industrial environments and defence supply chains, that is a continuity and compliance problem, not only a security one.

Trends: AI agents raise the stakes underneath the stack

Three developments are reshaping defence strategy at once. The first is autonomous agents. Enterprises are deploying AI agents that access data, make decisions and act at machine speed, and investment has followed: in the second quarter of 2026 alone, seed investors put around $360 million into securing and authenticating AI agents, and non-human identity was the most acquired category of the quarter (Omdia).

That work answers one question — what is this agent allowed to do. A second sits underneath it and is asked far less often: what is the agent running on, and is that still what it is supposed to be? An agent on a compromised host will follow every policy you gave it and still work for someone else.

Second, regulation is moving the burden of proof. NIS2 requires demonstrable continuity and incident management; the Cyber Resilience Act sets resilience requirements for products with digital elements themselves, not only the layers around them. Third, sovereignty is now an architectural requirement rather than a procurement preference.

Our approach: manage the system, not the attack

The industry has spent twenty years optimising the speed of a response that, for a quarter of real attacks, was never available at all. If defence depends on knowing what has to be fixed, unknown attacks cannot be pre-empted by definition. That is why we make the system itself the control, instead of the knowledge about the attack.

Our Pre-emptive Cyber Defense platform brings CyberSlice® technology together with CyOS™, our cyber-resilient computing foundation, into one security model. The operating environment an attacker would have to work in is continuously renewed and returned to a pre-validated, clean state, so previously used or compromised conditions do not persist by default — including when nobody observed that anything had happened.

What we design for is continuity, control and resilience: systems that keep operating during an incident, organisations that retain control of their own environment and data, and a smaller attack surface. No customer data has to leave the organisation for the security function to work.

From claims to measurement

Security claims are cheap, so we ask organisations to evaluate the technology under controlled conditions against criteria agreed in advance, with the customer measuring: recovery to a validated state, system integrity under attack, attack surface and exposed functionality, and continuity of trusted operation. The protocol and criteria are published up front, so results are reproducible and comparable.

Why Cybersec Netherlands matters

A claim like ours cannot be settled in a slide deck. It has to be examined by the people who run the systems in question, and that is what an event like this makes possible: practitioners, buyers and regulators in one building for two days, where the conversation around a demonstration reveals more than an RFP process does.

Where it happens matters too. Cyemptive is one of eight partners selected by HSD Security Delta for the HSD Meeting Zone, an ecosystem that has consistently moved European security innovation from research into deployment. We will show the technology in operation and how organisations can test it themselves. I present on 10 September from 16:15 to 16:45 in Masterclass Theatre 5.

Rob Pike is a Canadian citizen and the founder and CEO of Cyemptive Technologies. He will be at the Jaarbeurs in Utrecht on 9 and 10 September for interviews and background briefings.