SUNNYVALE, Calif., March 05, 2026 (GLOBE NEWSWIRE) — Cylake issued the following open letter today to leaders across the cyber community regarding the impact of generative AI on architectural approaches to security, and why it presents a fork in the road, rather than a dead end. To security leaders, CISOs, and operators navigating the AI transition, There is a growing narrative that AI will disrupt cybersecurity as we know it, and that generative AI systems will replace products, compress categories, and render large portions of the industry obsolete. I fundamentally disagree. AI will undoubtedly change cybersecurity, but it won't replace dedicated cyber platforms–and thinking otherwise could prove shortsighted. To understand what is happening now, we must rewind more than a decade. Machine learning has been part of cybersecurity for years, finding patterns in large datasets, building predictive models, and enabling faster remediation. Where humans struggled with scale, statistical models provided leverage. Generative models can now review code and infrastructure, identify anomalies, uncover vulnerabilities, and automate responses with increasing sophistication. The capability leap is real. However, GenAI is only as powerful as the data it can operate across, and this is where the industry faces an architectural constraint. For AI to deliver end-to-end protection, it requires comprehensive, unthrottled access to an organization's data, from network logs to endpoint telemetry. This data is orders of magnitude more than what most current products can ingest, process, and retain, because of the realities of cloud architecture constraints. There are two reasons for this. First, cloud economics can make it prohibitively expensive to ingest, process, and retain the volume of data needed for effective, GenAI-based security. The second, and perhaps more challenging, reason is that many organizations operate under regulatory, privacy, or security constraints that prevent data from passing through shared public infrastructure like the public cloud. At high levels of regulatory density and data sensitivity, multi-tenant cloud architectures make it structurally difficult to grant AI the unfettered access to complete security data it needs. This does not mean the cloud is broken for all cybersecurity use cases, but rather that a certain class of organizations needs a different approach than today's GenAI breakthroughs offer. I believe the market will segment, with more variety in security architectures. AI will change the way cybersecurity looks, but it won't replace it. In fact, I'm so certain of this that I just launched a new cybersecurity company. It may look like a risk in the current SaaS selloff inspired by AI, but over time, I'm confident it will prove to be the right choice. I urge any decision makers in the security space: now is the time to lean further into cyberprotection, instead of taking our foot off the pedal because “AI will replace everything.” GenAI may augment cybersecurity in some ways, but as an architectural examination shows, it isn't close to being a rip-and-replace solution for true protection. The only way AI kills cybersecurity is if we let it lull us into accepting less than purpose-built protection. Respectfully, Media contact:
Nir Zuk
Founder and CEO
Cylake
cylake@bulleitgroup.com
