Each August, the cybersecurity industry converges on Las Vegas for Black Hat USA, one of the year’s largest gatherings of security researchers, CISOs, technology vendors, government officials and enterprise practitioners. The conference is known for technical research presentations, product announcements and discussions around the threats and technologies shaping cybersecurity strategy. For enterprise leaders, it offers a snapshot of the issues commanding the industry’s attention — and often where security investment is headed next.
Unsurprisingly, AI is the defining theme of Black Hat USA 2026, officially and unofficially; this year’s event includes a dedicated AI Summit and a new AI Zone, alongside keynote discussions focused on AI-powered cyber operations and the changing dynamics between attackers and defenders.
“AI is everywhere; from signage when you land at the airport, to partner booths and the start of every client conversation,” said Julie Talbot-Hubbard, vice president of security services at enterprise IT consulting and services firm AHEAD.
The volume of AI messaging has also created a new challenge for security leaders: separating meaningful advances from marketing claims.
Chase Cunningham, chief strategy officer at software virtualization platform Demo-Force, described AI as having “annexed the conference.”
“You cannot walk 20 feet without encountering agentic, AI-powered or autonomous attached to a product that, in some cases, was apparently doing just fine without those words last year,” he said.
AI as a tool, not a savior
Despite the flood of AI-related announcements, attendees described a shift in tone: less focus on AI’s potential and more on the operational challenges of deploying it securely. “The conversation feels more mature this year,” as Talbot-Hubbard put it. More pragmatic.
Rather than debating what AI might eventually enable several years down the line, security leaders were focused on how to identify, govern and secure the AI systems currently entering the enterprise.
“People are moving past the hype and focusing on the hard operational questions around securing agents, identities, permissions and the infrastructure that supports them,” said Diana Kelley, CISO at Noma Security, an agentic AI security platform.
Talbot-Hubbard agreed, saying AI is now “part of the broader security operating model conversation, not a separate experiment.” Organizations are looking beyond individual products, to whether they have the foundational capabilities needed to support AI securely, including identity controls, visibility, governance and resilience.
“The most positive surprise was how practical the AI conversation has become,” Kelley said. “There’s much less patience for generic ‘AI-powered’ claims and much more focus on provable controls, observability, governance and whether these products actually work in complex operational environments.”
For some attendees, the sheer number of AI-focused companies at Black Hat highlighted both the level of investment flowing into the space and the difficulty of evaluating which technologies will have a lasting impact.
“The number of new, well-funded AI-first companies on the floor with big booths [and] serious presence” was notable, said Louis-David Mangin, CEO and co-founder of Confiant, a cybersecurity solution for advertising. “It’s a signal of how much capital is chasing this space, but it also creates noise at an event that used to be easier to navigate.”
Mangin said the increased commercial presence also changed the feel of the conference compared with previous years. While conversations about real challenges and hard-won lessons continue, he noted they can sometimes feel harder to find amid the volume of vendor messaging.
AI raises the stakes for familiar security challenges
In those conversations, attendees repeatedly returned to longstanding security challenges that remain unresolved even as AI is introduced. AI is both a security tool and an accelerant for attackers; the concern is less that cybercriminals will develop entirely new techniques, and more that AI will allow them to operate more efficiently.
“The overarching concern is that offense is getting cheaper, faster and easier to scale,” Cunningham said. AI can accelerate reconnaissance, vulnerability analysis, exploit development, social engineering, credential abuse and post-compromise activity, he said.
Kelley similarly pointed to the speed of attacks as a major concern, noting that AI is compressing the time between discovery and exploitation while organizations continue to manage existing security debt, including overprivileged identities and software supply chain exposure.
The greater capabilities of cybercriminals are also emerging as enterprise architectures grow more complex and integrated. Mangin described a growing sense among attendees that organizations are increasingly vulnerable as a result.
“The attack surface isn’t just bigger, it’s more interconnected,” he said, pointing to risks spanning enterprises, platforms, AI systems and infrastructure. “Attackers are looking for the gaps between them.”
That interconnectedness is also why security leaders continue to emphasize core practices. AI may accelerate attacks, but it does not eliminate the need for strong identity controls, least-privilege access, monitoring, segmentation and recovery planning.
“AI is an accelerant, not an exemption from doing the fundamentals,” Cunningham said.
The challenge for enterprises
Across discussions at Black Hat, one theme emerged consistently: organizations are trying to move quickly enough to capture AI’s benefits while avoiding the risks created by deploying new capabilities without sufficient controls.
Mangin described this as an ongoing imbalance between attackers and defenders. Cybercriminals can experiment and adapt without the same policies, processes and governance requirements that enterprises must navigate. This naturally puts enterprises on the back foot and places greater pressure on the security decisions they do make.
This challenge is what motivated many IT leaders to attend Black Hat USA 2026, where they can converse with their peers and learn new ideas for how to get ahead of the threat actors.
Talbot-Hubbard spoke of wanting a sharper read on security leaders’ priorities for the next 12-24 months, as well as the gap between strategy and execution — but she also emphasized the importance of connecting first-hand with the industry to compare notes: “The best outcome isn’t a list of new technologies; it’s insights that help organizations make better decisions with more confidence.”
Other attendees see Black Hat as an opportunity to gain greater oversight of an increasingly complex picture.
“We came here to deepen and widen our view,” Mangin said. “No one company sees the entire picture. Everyone has a different piece of the puzzle. Black Hat gives us a chance to understand what’s happening beyond the part of that chain we see directly and learn from people with visibility into other parts of it.”
And then there’s the desire for clarity. Cunningham said he came to Black Hat to distinguish those who have built tools that deliver measurable security outcomes from those that simply put the word agentic on last year’s architecture diagram. He also wanted to speak with practitioners dealing with issues in real environments, challenge vendors on product capabilities, and understand where research is moving faster than enterprise security programs can adapt.
Simply put, he said: “I am here to separate signal from noise.”
Were you at Black Hat this week? Let us know what stood out for you: [email protected].
