Zscaler Accelerates AI Security Push as Anthropic Threat Model Drives Zero Trust Urgency
June 9, 2026 - Zenith Live Conference
Zscaler is experiencing an unprecedented level of board-level urgency around AI security following Anthropic's Project Methos, with CEO Jay Chaudhry characterizing the current moment as potentially larger than COVID for cybersecurity demand. During the company's Zenith Live conference, Chaudhry revealed that nearly every CIO he has spoken with in recent months has established task forces reporting to their boards weekly or biweekly on progress, a level of executive engagement that exceeds what the company saw during the pandemic.
The company disclosed that its AI Protect solution, launched just in January, has already crossed over $100 million in trailing twelve-month bookings. More significantly, Zscaler announced Zero Trust for AI Agents, which Chaudhry positioned as potentially the company's largest transaction-based monetization opportunity, with agent-related transactions potentially requiring "a couple of zeros" added to the company's current 750 billion daily transactions.
Methos Threat Model Creates Secular Shift to Zero Trust
The Anthropic Methos vulnerability scanning model is fundamentally changing customer behavior in ways that directly benefit Zscaler's architecture. Chaudhry explained that Methos and similar AI-powered vulnerability scanners from OpenAI are discovering vulnerabilities at extraordinary speed, creating an environment where traditional patching approaches will never keep pace. "The answer is not that you're going to double, triple down on just patching. You will never get out of doing patching itself," Chaudhry stated.
This reality is driving customers toward two key defensive strategies where Zscaler excels: hiding attack surfaces through application concealment and preventing lateral movement through microsegmentation. Wayne Fudersky, Deputy CISO at Edward Jones, noted during a customer panel that financial services regulators have already increased their compliance questions around AI and are challenging firms on how they control and manage exposure. "We're being challenged constantly on how are you handling this," he said.
The urgency is manifesting in concrete ways. Chaudhry reported that large enterprise customers who previously implemented only partial Zero Trust Private Access deployments are now moving to full ZPA coverage for all users. The company saw non-seat-based revenue grow from approximately 25% of new ACV a quarter ago to 30% in Q3, indicating accelerating adoption of infrastructure-based products like Zero Trust Branch and Zero Trust Cloud workloads.
Zero Trust for AI Agents Addresses Critical Governance Gap
Zscaler's announcement of Zero Trust for AI Agents represents a fundamental architectural approach to what customers describe as their number one barrier to production AI deployment. "The number one message that comes from CIOs and CSOs is we have identified a few pilot programs, we are ready to roll out, we have built some agents, but I'm uncomfortable because the governance and controls aren't there," Chaudhry explained.
The solution leverages approximately 70% of Zscaler's existing Zero Trust Exchange infrastructure but required new AI brokers for MCP protocols and A2A communication, along with capabilities to extract prompts and analyze intent and risk. The company is positioning this as an in-line policy enforcement mechanism that addresses a fundamental problem with traditional application-based governance: AI agents can bypass applications entirely and go directly to data sources, eliminating traditional control points.
Mustafa Keeve, Chief Security Officer at UKG, articulated the market need clearly: "I can't secure what I cannot govern. I cannot enforce policy if I don't know what's going on." He noted that UKG is actively implementing Zscaler's prompt inspection and governance capabilities, calling the maturity progression over the past 18 months "a game changer."
Strategic Acquisitions Create Differentiated AI Security Platform
Zscaler's acquisition strategy is proving particularly well-timed for the AI security moment. The company's purchase of SPLX brought automated red teaming capabilities that can continuously test AI models using any underlying scanning models, including Methos. Keeve, who was a SPLX customer prior to acquisition, emphasized that automated testing across security, sentiment, and validation was critical for UKG's plan to deploy AI agents to its 80,000 customers.
The Symmetry Systems acquisition is enabling what Zscaler calls Access Graph, which Chaudhry described as solving "a very hard problem" that creates significant barriers to entry. The technology aggregates metadata from applications across an enterprise to understand which identities and entities are accessing which data sources, creating what the company calls a context graph. This becomes essential not just for data governance and SOX compliance in the traditional sense, but for applying policies in the Agentic Exchange to determine which groups of agents can access which applications or data sources.
CFO Kevin Rubin noted that Red Canary is being integrated into a new Agentic SecOps offering that the company formally launched at the conference, with Adam Geller leading the initiative. The approach leverages Zscaler's position as a center of gravity for data, with detections and investigations powered by Agentic frameworks across ZIA, ZPA, DLP and client data that previously would have only been available by streaming to third-party SIEM platforms.
Serviceable Addressable Market Expands to $120 Billion
Zscaler now estimates its serviceable addressable market at over $120 billion, with meaningful expansion coming from newer categories. The company breaks this into four segments: Zero Trust Everywhere at approximately $65 billion covering users, cloud and branches; Data Security as an ever-growing market accelerated by AI-driven data creation; Agentic Ops at approximately $30 billion primarily around SecOps and IT operations in early stages; and Securing AI as a brand new market segment where the company has established what Chaudhry characterized as "serious momentum."
The Zero Trust Branch opportunity is gaining particular traction. Jason Coler, Deputy CISO at Eaton, explained that the manufacturing company has deployed Zero Trust appliances in approximately 100 factories specifically because SD-WAN couldn't provide security down to the device level. "We can't have downtime. Our production and our plants need to keep running," he stated, adding that the ability to isolate and contain potential incidents was critical.
Notably, over 600 enterprise customers now have Zero Trust Everywhere deployments, meaning they have implemented Zero Trust for users, branches, and cloud workloads. This represented growth from 550 customers disclosed just one quarter ago, indicating accelerating adoption of the full platform approach.
Go-to-Market Adjustments Target Under-Penetrated Segments
President Mike Rich addressed investor concerns about Q3 guidance, which had cited a couple of key sales departures as contributing to a more conservative outlook. Rich characterized this as "just two at the same time" and emphasized that strong performers see clear career paths at Zscaler, particularly as the company promotes from within when departures occur. He noted that turnover can be healthy for raising the game of the broader organization.
More significantly, the company is rebalancing its go-to-market focus toward the 2,000 to 10,000 employee segment, which falls within Zscaler's target universe of approximately 20,000 large enterprises but has historically received limited coverage. "Everybody wants to work on existing customers because it's just easier to get deals done," Rich acknowledged. The company brought in outside analysis and realized "this is a big opportunity" requiring more focused territory planning and alignment with marketing spending.
Rich emphasized that large strategic deals continue to progress, though testing cycles can extend as enterprises validate configurations and integrate multiple stakeholder groups. "Sometimes there's new requirements they want to see us deliver before they're ready to take that deal to the next level," he noted, adding that Zscaler's close customer relationships and rapid feature delivery remain differentiators.
Pricing Evolution Toward Consumption-Based Models
The company provided updated disclosure on its shift away from pure seat-based pricing. Non-seat-based new ACV reached 30% in Q3, up from 25% disclosed the prior quarter. Zero Trust Branch pricing is based on number of devices and traffic flows, Zero Trust Cloud on workload count and traffic, and the new Agentic Exchange will be based on agent counts and traffic, which translates to token consumption.
When asked about customer comfort with token-based pricing given the uncertainty around AI costs, Edward Jones' Fudersky provided a measured response: "Comfort is probably a strong word. I think there's an expected cost, but comfort is something I think we just have to get better at." He drew parallels to early cloud adoption when enterprises "just liked to throw everything out there" before learning to manage expenses, suggesting a similar learning curve lies ahead for AI-related consumption charges.
Chaudhry emphasized that Zscaler is "still figuring out" exact pricing for the Agentic Exchange, working with the first dozen or two customers to understand traffic flows before firming up pricing models over the next couple of months. The company expects to share progress metrics as deployment scales.
Product Velocity Accelerates Across Platform
Beyond AI security, Zscaler continues delivering approximately 200-plus features annually across its ZIA and ZPA products. Recent enhancements include browser extension capabilities through the SquareX acquisition, B2B exchange for supply chain security that competes primarily against 30-year-old site-to-site VPNs, and power segmentation improvements. The company introduced a unified agent framework enabling Agentic interfaces across all products, with ZDX agents for automated digital experience management serving as an early example.
In AI Protect specifically, the company announced support for over 250 generative AI applications with full prompt inspection, up from a smaller set at launch. The solution now provides visibility into embedded AI within SaaS applications and AI activity on endpoints including permissions and risk analysis. Integration with Anthropic and OpenAI compliance APIs is complete, and the company added standalone prompt hardening features and red teaming for MCP servers as those protocols gain adoption.
Chaudhry noted that Zscaler is approaching AI security "like a startup within Zscaler," with dedicated resources under Chief Product Officer Dhawal Sharma building products at extraordinary velocity. "I couldn't be happier with the pace at which we are building and developing these products and the traction and interest we're drawing from customers," he stated.
Competitive Positioning Strengthens in Zero Trust Architecture
Chaudhry drew sharp distinctions between Zscaler's architecture and traditional security approaches. "When others talk about SaaS, they talk about secure access to users" in the area Zscaler pioneered, he noted, but competitors implementing security in the cloud are "spinning up virtual firewalls fundamentally. There's no Zero Trust in it." He argued that firewall vendors cannot pursue true Zero Trust because "it cannibalizes all the firewalls" and cited the parallel to telecom providers initially fighting MPLS elimination.
On identity, Chaudhry clarified that Zscaler does not compete with basic identity provision from vendors like Okta or Microsoft, but rather adds authorization services on top. "Identity gets created, we take that, we become the switchman, but now we can add a number of authorization services on top of that: skills, tools, access, all the other stuff," he explained. For agents specifically, these additional authorization layers become "extremely important" beyond basic identity, which is why the Symmetry Systems Access Graph technology matters for understanding data lineage and governance.
The company's global infrastructure of 160-plus exchanges, including public and private deployments for major customers, creates barriers to entry that AI models cannot replicate. "An agent is not going to go and create all of the infrastructure for your connection, network traffic routing, all that stuff is a fairly complex and sophisticated area," Chaudhry noted, arguing that increased agent deployment will drive more demand for policy enforcement and in-line inspection rather than displacing security providers.