LAS VEGAS – CrowdStrike CEO George Kurtz is pinning the company’s future on what he calls the “agentic SOC” – a security operations centre powered by fleets of AI agents designed to keep pace with adversaries moving faster than ever.
Speaking in a wide-ranging interview with SiliconANGLE post his keynote address at the company’s annual event, Fal.Con 2025, Kurtz said that the accelerating speed of cyberattacks has left traditional SOCs overwhelmed.
“It used to be weeks, then days, then hours and minutes, now and seconds,” he said. “The SOC analyst is overloaded with alerts. There is too much coming in, and there is not enough time to get through this at the speed at which the adversary is operating.”
The company’s answer is an autonomous layer of AI assistants capable of performing much of the investigative and remediation work on behalf of human analysts. Kurtz envisions a future where analysts supervise a fleet of AI agents, comparing it to the way one operator can manage multiple self-driving cars.
“Our big announcement was Agent Works, which allows customers to build their own AI agents to get ahead of threats,” he said. “One day we are going to have an autonomous SOC analyst doing the work of the analyst, controlled by the human as a fleet of agents.”
Building security AGI
Kurtz places this vision in the broader context of what he calls “security AGI” – an eventual stage where AI can continuously learn and operate beyond human capacity. He drew a parallel to the five levels of vehicle autonomy, saying CrowdStrike is working towards level five for security analysts.
“Ultimately that is going to be security AGI,” he said. “It will be self-operating, continuously learning and fully autonomous. And we want to get there first.”
CrowdStrike’s approach relies heavily on data. The firm has accumulated what Kurtz described as “a treasure trove of security data” over its 14 years, encompassing trillions of daily telemetry events from customer sensors, annotated threat data from its managed detection services, and intelligence from incident response work.
“You cannot build large language models without the correct data,” he said. “Our data is enriched with the customer’s own information, and now they can build in one spot with all the relevant data on the CrowdStrike Falcon platform.”
Managing fleets of agents
Kurtz argued that AI agents need the same kind of monitoring and guardrails as human employees. Each agent, he said, will have an identity, access to workflows, and the ability to interact with systems outside a corporate network.
“How we manage that is really the first AI-DR type technology,” he said, referring to detection and response. “From a compliance standpoint, I do not think people realise how much compliance is going to drive the protection around AI agents.”
Some customers are already assigning employee numbers to AI agents, he said, underscoring how integrated they are becoming in daily workflows. “They are going to be like digital employees working 24×7, and analysts will be managing them like a team.”
Pangea deal
CrowdStrike’s acquisition of AI security company Pangea earlier this month was part of this strategy. Pangea develops technology to protect AI agents at the “prompt layer” – defending against attacks such as prompt injection and model jailbreaks.
“It is fantastic technology,” Kurtz said. “They built for developers as well as consumers of AI technology. You want to prevent as much as you can upstream, and they do that.”
Unlike some rivals that pursue large-scale acquisitions, Kurtz said CrowdStrike focuses on smaller deals that can be fully integrated into its Falcon platform. “Our philosophy is long-term thinking,” he said. “Customers don’t want four platforms. Ultimately, when companies acquire too much, it is digital taxidermy – it looks alive, but it’s Frankenstein underneath.”
Platform and licensing
Since its founding in 2011, CrowdStrike has expanded from a single module to 30. Kurtz said customers are consolidating their security stacks around a smaller number of suppliers, and CrowdStrike’s “Falcon Flex” licensing model has accelerated adoption.
Flex allows customers to commit spending up front and then draw on the entire product catalogue without separate procurement cycles. Kurtz said the model has over $3 billion in total contract value. “We make it simple for customers to buy, consolidate and expand on our platform,” he said.
The approach has led customers to immediately ask about new acquisitions like Pangea, he added. “As soon as the deal is closed, it’ll be on the price book. You don’t have to wait months.”
Asked who CrowdStrike is competing against, Kurtz said the company’s priority is protecting customers rather than beating rivals. But he acknowledged that the race to security AGI is real. “If we are the first to provide a fully autonomous SOC analyst, that is the win,” he said. “We have got the right data set, the right people, the right focus.”
Kurtz rejected the idea that CrowdStrike is moving from product to platform only recently. “We have been a platform for a long time,” he said. “There are only a few platform players in security, and we are one of the big ones.”
Public markets and long-term outlook
Kurtz also addressed the challenges of being a listed company in the United States. Some policymakers have floated reducing reporting requirements from quarterly to twice a year. “We will do whatever regulatory requirements are,” he said, while stressing that CrowdStrike is managed for long-term growth rather than short-term targets.
“A lot of decisions can be made for the short term that look good for 90 days, but we won’t do that,” he said. “We are focused on being a generational security company.”
CrowdStrike, which reported revenue of about $4.5 billion, generates cash flows that Kurtz described as strong for its size. “We are going to add gas and continue to grow,” he said.
He argued that AI will expand the total addressable market significantly. “If we made a big business protecting every PC, we are going to make an even bigger business protecting every AI agent.”
Future of work and security
For analysts worried about automation replacing them, Kurtz offered reassurance that AI will augment rather than eliminate their roles. “Sometimes people get scared about AI. The reality is, it is going to allow people to do more and faster,” he said.
He imagines a day when SOC analysts start their shifts by checking in on digital employees that have been working overnight. “They are going to manage a fleet of agents, and if they need another agent, they will create it. They will work together almost like a mesh network,” he said.
Kurtz is confident that as technology evolves, demand for security will grow alongside it. “Security parallels the slope of the technology curve,” he said, adding that “it does not matter what technology does, there is always going to be a play for security. We win as long as technology evolves.”
