In 2026, it’s clear that cyber risk isn’t coming from one major new threat. It’s coming from lots of different ones adding up. More apps. More identities. More suppliers. More automation. And more AI quietly doing work in the background. Most organizations are moving faster than their ability to see who has access, what’s trusted, and what’s acting on their behalf.
Malefactors are taking advantage of that gap. They’re abusing tokens instead of passwords, exploiting on supply chains instead of direct breaches, and using automation and AI just as comfortably as defenders. At the same time, regulators, insurers, and boards are asking harder questions. They want evidence, not promises.
This 2026 Cyber Predictions series brings together perspectives from across the industry on what’s really changing, what’s already breaking, and where businesses need to focus next. Less hype and a whole lot more reality.
Let’s see what the experts have to say.
Token Hopping, Reverse Ransom
Rik Ferguson, Vice President of Security Intelligence at Forescout, has two cyber trends for 2026 that he thinks are going to make the biggest impact. Token hopping and reverse ransom.
“Token hopping is what happens when attackers stop chasing passwords and start stealing trust. OAuth tokens, app consents, refresh tokens. The stuff that lets a malicious actor move through your SaaS as if they’re a legitimate user and sometimes persist even after a password reset. If you don’t have a clean view of which apps are authorized to act on behalf of users, you’re not managing risk, you’re guessing.”
Ferguson describes reverse ransom as extortion with leverage. “Compromise a smaller supplier, disrupt the chain, and put pressure on bigger downstream organizations to pay. Not because they were hacked directly, but because operations are at a standstill. And of course, the rest continues to accelerate. AI-driven social engineering kits will be delivered as a service in 2026.”
He adds that the exploitation of edge and IoT vulnerabilities is gaining even more traction. “Specialist cybercrime supply chains, hacktivist disruption of OT, and the slow but real push toward quantum readiness. How do we face up to this? Inventory and restrict app consents, revoke stale tokens, map supplier dependencies, and rehearse supplier down response like it’s a fire drill”
A Shifting Regulatory Landscape
Jamie Akhtar, CEO and Co-Founder of CyberSmart, says: “The cyber market and its regulatory landscape are shifting quickly and organizations are starting to feel the pressure of a more demanding regime. This will continue throughout 2026. As the Cyber Resilience Bill comes into force, it brings with it mandatory adoption of the Cyber Assessment Framework across critical sectors. The scope of regulation expands as the definition of Relevant Managed Service Providers is broadened, placing managed service providers (MSPs) directly in the regulatory spotlight. This change introduces new duties around incident reporting, baseline security controls and formal assurance, meaning that both service providers and their customers must operate with far greater transparency and discipline.”
He says the CyberSmart 2025 MSP survey saw that this was already starting to happen. More than three-quarters (77%) of MSPs reported that their businesses’ security capabilities were already coming under greater scrutiny by prospects and customers. “This suggests that MSP customers are more aware than ever of the importance of good cyber credentials in a potential partner – and this will only continue.
Insurers, meanwhile, are moving away from the old model of static cyber questionnaires, which no longer offer adequate insight in a climate of rising breach frequency and cost. For small and medium-sized businesses (SMBs) in particular, insurers are increasingly requesting continuous security telemetry as a prerequisite for underwriting. This shift signals a market-wide recognition that real-time visibility is the only reliable basis for evaluating cyber risk.”
At board level, expectations are also rising, Akhtar adds. “Directors are no longer satisfied with policies that look impressive on paper but cannot be demonstrated in practice. They want clear, verifiable evidence of operational cyber hygiene, which is accelerating the use of automated evidence collection and continuous control monitoring. The demand is for assurance that is lived day to day, not compiled once a year.
Supply chain risk has become equally hard to ignore. High-profile interventions such as the FTSE 350 cyber letter and the latest CSM v4 requirements for defense suppliers have pushed the issue into the mainstream. Large organizations now expect their upstream suppliers, including SMEs, to show that they have implemented basic controls and maintain resilience in a consistent and certifiable way. The cumulative effect is a market that values demonstrable, continuous cyber competence over declarations of intent.”
Lingering Blind Spots
James Moore, Founder & CEO of CultureAI, adds that as we move into 2026, the biggest risk isn’t AI itself, rather it’s the blind spots organizations still have around how their people, and their tools are actually using it. “Almost everybody is now using AI platforms, often without knowing what data those tools retain or how it’s used. With an abundance of AI comes an abundance of data loss. I predict three major threat shifts that will define 2026:
1. The rise of invisible AI usage, especially in everyday SaaS
What people think of as ‘AI tools’ is too narrow. An AI app is any SaaS application that takes data and passes it into a model. Most organizations haven’t even scratched the surface of understanding that. I believe that embedded AI features within SaaS apps, beyond common AI tools like ChatGPT or Copilot, could contribute to enterprise data-loss incidents next year.
2. Legacy controls will continue to fail, not because they’re bad, but because they weren’t built for this problem
To solve AI data-loss, you have to understand the contents of every request going to an AI app. DLPs and CASBs simply weren’t built for that. You can’t just turn those apps off and block them all and hope for the best.
3. Agentic AI will create a new class of blind spots
I expect that we will see the emergence of AI agents that act, browse, and make API calls independently. When AI starts taking actions on your behalf, you move from securing human behaviour to securing autonomous behaviour. Most organizations aren’t remotely ready for that.”
However, Moore also believes that 2026 will be the year that enterprises unlock AI at scale. “However, this can only be done if they treat usage as a governance and enablement problem, not a blocking problem. Our job isn’t to scare people away from AI. It’s to give them the visibility and control to use it safely, at speed. The organizations that win in 2026 will be the ones that move to the top-right quadrant: high adoption and high security, not one or the other.”
An Identity Explosion
Mark McClain, CEO and Founder at SailPoint says if AI agents are left unchecked, enterprises could face an identity explosion. “Next year, businesses will need to get a handle on their AI agents if they want to ensure the risks don’t outweigh the benefits. Agentic AI isn’t the future – the new digital workforce has already arrived, with 82% of businesses currently utilising AI agents.
“Despite widespread adoption, less than four in ten organizations are currently governing AI agents. If enterprises aren’t careful, they could face an ‘identity explosion’, where unchecked AI agents introduce vulnerabilities that spiral outwards from within the business itself, triggering compliance violations or inadvertently exposing customer information.
“To keep agents in check, organizations will need to approach AI agent access rights in the same way they would humans. That means that previously tried-and-tested, static approaches to access policies just won’t cut it in 2026.
“Businesses need a new model for identity security: one that is adaptive and contextual and supports the adoption of AI agents in a secure, scalable manner. Next-gen identity security tools will be a crucial consideration for businesses looking to roll out contextual, precise, real-time access control policies, where access is purposefully granted when appropriate – and aggressively revoked when not.”
From Technologist to Strategist
Rex Booth, CISO at SailPoint believes 2026 will see CISOs become great strategists, not just technologists.
“Recently, we saw Shiny Hunters borrow social engineering tactics from Scattered Spider and it’s not an isolated incident. Gangs trade knowledge, tactics, tools and even people. Plus, ransomware-as-a service has erased many technical barriers and made cybercrime accessible to anyone with time, a laptop, and an internet connection. Crime now moves faster, enabled by easy access to knowledge and capabilities.”
Booth says all this sharing means two things in the year ahead. “Attacks are going to get more frequent, and the results will be less predictable. CISOs looking to stay one step ahead need to be great strategists, not just technologists. Keeping crime out means securing buy-in from the wider business – getting them to view security as an enablement function. Traditionally, security has been viewed as the department of ‘no’, but we’re not just here to block things. If we’re going to keep things safe in 2026, we need our stakeholders to understand we’re collaborators, not obstacles.”
Treat AI Competence as a Learnable Skill
Organizations need to treat AI competence as a learnable skill, says Joel Carusone, SVP of Data and AI at NinjaOne. “In 2026, workplace culture will increasingly be shaped by how organizations respond to the growing gap between AI adoption and formal education. AI is already a part of daily work, while broad curriculum reform on the topic is still several years away. As a result, many employers will need to decide how much responsibility they take for developing AI skills within their own teams.
“Prompt engineering, output validation, and critical thinking about AI are becoming essential skills. In many organizations, these capabilities are taken for granted, rather than formally taught, leaving employees to learn on their own. Without clear guidance, this can lead to inconsistent outcomes and varying levels of confidence for employees using AI at work.
Carusone adds that more resilient organizations are likely to take a more deliberate approach. “By treating AI competence as a learnable skill, they can provide practical training, set clear expectations for responsible use, and reinforce the importance of human oversight. This helps teams use AI more effectively while maintaining trust and quality.
“In 2026, the key workplace culture difference won’t simply be whether companies use AI, but how intentionally they support their people in developing the skills they need to use AI to their advantage.”
Information Security Buzz News Editor
Kirsten Doyle has been in the technology journalism and editing space for nearly 24 years, during which time she has developed a great love for all aspects of technology, as well as words themselves. Her experience spans B2B tech, with a lot of focus on cybersecurity, cloud, enterprise, digital transformation, and data centre. Her specialties are in news, thought leadership, features, white papers, and PR writing, and she is an experienced editor for both print and online publications.
The opinions expressed in this post belong to the individual contributors and do not necessarily reflect the views of Information Security Buzz.