At Palo Alto Networks, we are privileged to sit at the intersection of cybersecurity technology, business, and outcomes. We speak to hundreds of customers, partners, analysts, and other experts every day about threats, tools, and solutions, so we’re often asked for our opinions on what’s next for cybersecurity. But when technology leaps forward, it can create trends whose impact and speed are much harder to forecast. The sudden changes to the cybersecurity landscape in the first half of 2026 are a good example of this, leaving many playing catch-up in some areas. To illustrate the salience of these changes, let’s take a look back at our predictions for this year, as outlined by our Chief Security Intelligence Officer, Wendi Whitmore, in her late-2025 article ‘2026: The Year of the Defender’, and see which have come true, which haven’t, what’s surprised us, and what’s next. The new age of deception: The threat of AI identity The role of frontier AI in shaping cyber risk and vulnerabilities has become far more significant in 2026, impacting many of our predictions. AI identity is just one significant part of that.The transition from human identities to non-human identities (NHIs) is not only well underway but progressing even faster than we predicted in late 2025. The ratio of NHIs to human identities was 82:1, now it’s a staggering 109:1. We’ve all had to step up our capabilities in pressure testing identity systems to simulate bad actors’ increasing use of NHIs in their attacks. It’s a trend our CyberArk acquisition did anticipate, however, leading us to build our new Idira platform around identity. Thankfully, our access to new frontier AI models via our participation in Project Glasswing has allowed us to model some of these emerging attack vectors. It’s a good thing we can, because in the