AGENTIC AI : AGENTS CONTROL AT STAKE
Recent reports show that, in the context of expanding agentic AI, Nonhuman identities have quietly overtaken humans in most environments. Yet most security architectures treat agents like an afterthought if they register them at all. Warnings surface that without strong verification, delegated access and life cycle governance, autonomous agents could become the next major blind spot in enterprise defence. “We like to say that identity access management is really table stakes for any agentic AI implementation,” US expert Preis says. “Agents themselves have an identity. That identity has to be not just authenticated, but verified.” In fact, security teams face a harder question than simply how to authenticate them: how to prove which agent took an action, to establish the boundaries of its authority, and to stop it when something goes wrong?
PHP GIVES NOTICE OF DEFAULTS
The software vendor PHP, a popular general-purpose scripting language that powers blogs and websites, announced on October 21 having detected two painful vulnerabilities. A patch was promised for the same day.However, old unmaintained versions must be upgraded before patching.
DUTCH FRENCH PARTNERSHIP FOR INNOVATION AND TECHNOLOGY
Important meetings were held during September and October 2026 in the context of this ongoing agreement. Among others, Cybersecurity was on the agenda. The ambition: de-risk our strategic dependencies, aggregate demand and accelerate procurement and market adoption of Dutch and French cybersecurity solutions. Even more encouraging: France and the Netherlands will now have a dedicated sub-focus group on Cybersecurity and AI, turning shared ambition into concrete Franco-Dutch cooperation. We will enlarge this first ‘starters group’ gradually, with more interested and relevant organisations.
DUTCH START UP PASSGUARD MAKES IT CLEAR
Our pleasure to reproduce a Post by Tom Leitje, CEO of PASSGUARD, a Dutch startup expert in darkweb monitoring and infostealers detection. No need to say Tom’ vision is totally shared by the ECA.
We turned down tens of thousands of dollars in free hyperscaler credits. And to be completely honest, when we first received those offers, we were practically drooling over them. I mean, tens of thousands of dollars in free cloud credits? As a startup, just imagine what you could do with that!
But the more we thought about it, the more it started to feel like a deal with the devil.
A faustian pact.
Because here’s the catch: The deeper you integrate into a hyperscaler’s ecosystem, the harder it becomes to leave. And once those free credits run out, you’re left with an infrastructure that’s increasingly difficult to migrate, while your cloud bill can grow rapidly as your usage increases.
Suddenly, those free credits don’t seem quite so free anymore.
And then there’s another consideration: How can we claim to be building the European alternative for dark web monitoring and infostealer intelligence, while making ourselves increasingly dependent on American hyperscalers?
So we decided to do things differently. Earlier this year, we moved Passguard’s core infrastructure away from Microsoft Azure to six European providers. Was it the easiest option? No. Was it the cheapest in the short term? Definitely not. But it gave us more control over our infrastructure, reduces our dependence on individual suppliers, and allows us to support the European technology ecosystem we’re trying to be part of.
You can listen to the full podcast or read the accompanying article (both in Dutch):
EUROPEAN FUNDED CYBERSECURITY PROJECT
A new ECCC funding opportunity is now open for applications!
The call Cyber-12 under the Digital Europe Programme focuses on strengthening the 𝐧𝐚𝐭𝐢𝐨𝐧𝐚𝐥 𝐚𝐧𝐝 𝐜𝐫𝐨𝐬𝐬-𝐛𝐨𝐫𝐝𝐞𝐫 𝐜𝐲𝐛𝐞𝐫 𝐡𝐮𝐛𝐬 across the 🇪🇺.
With this action, the ECCC contributes to the implementation of the Cyber Solidarity Act, which provides for the establishment of a pan-European network of cyber hubs as part of the European Cybersecurity Alert System. By leveraging advanced technologies, these hubs will help strengthen the detection, analysis and response to cyber threats.
𝐁𝐮𝐝𝐠𝐞𝐭: EUR 10 million
𝐃𝐞𝐚𝐝𝐥𝐢𝐧𝐞: 16 March 2027
This funding mechanism includes two interrelated calls: a call for expression of interest and a call for proposals. Each applicant will have to apply to both workstreams.
