Baird: Zscaler Bets the Agentic Era Will Dwarf Its Entire Existing Business — and It Has a Plan to Monetize It
Baird Global Consumer Conference, June 2, 2026 — CFO Kevin Rubin lays out the next phase of the Zero Trust platform
Zscaler CFO Kevin Rubin used a fireside chat at the Baird Global Consumer Conference to make one of the most consequential forward-looking statements the company has offered in some time: the volume of agent-to-agent and machine-to-machine traffic is expected to be "orders of magnitude greater" than the over 50 million human users currently running through the Zero Trust Exchange. If that projection is anywhere close to correct, the revenue opportunity from the agentic era alone could exceed everything Zscaler has built to date — though the monetization model is still being assembled in real time.
The Agentic Exchange: The Biggest Opportunity Not Yet in Market
The most important thing investors need to understand is that the agentic exchange — Zscaler's framework for applying Zero Trust principles to agent-to-agent and machine-to-machine communication — is not yet commercially available. Rubin was direct about this, noting that "the agentic exchange is not actually in market yet," but confirmed that the architecture mirrors what the company has already proven at scale with users, devices, and workloads. The monetization model will likely be consumption and token-based rather than seat-based, consistent with the broader direction the business is already moving.
To accelerate its positioning in this space, Zscaler recently acquired Symmetry, a company that has built an access graph designed to infer inherited permissions flowing through AI agents. Rubin described it as "another piece of the agentic exchange puzzle," giving Zscaler the ability to understand agent identity and permissioning within its inline policy enforcement layer. The strategic logic is clear even if the financial contribution is still minimal.
Mythos Changed the Customer Conversation Overnight
Anthropic's Mythos model announcement has had a measurable effect on how Zscaler is received in enterprise security conversations. Rubin explained that Mythos and similar frontier models are now identifying vulnerabilities "at a rate and pace that far exceeds what we've seen today," creating a backlog problem that already-stretched IT teams simply cannot manage. "Large organizations today already have a backlog of patches for existing known vulnerabilities that they are incapable of handling in any reasonable period of time," he said. "What Mythos and other models are identifying is a volume of vulnerabilities that already exceeds what they can handle today — you're just piling on top an order of magnitude more."
Zscaler's answer is structurally different from patch management: hide the applications entirely so they cannot be reached, and enforce least-privilege access for every session. "If you cannot reach the application, you cannot breach it," Rubin said. The argument is that the AI-driven vulnerability explosion makes Zero Trust not just preferable but practically necessary, and that inbound customer interest has increased materially since Mythos became public. This is a genuine demand catalyst, not a talking point.
AI Protect at $100 Million in Bookings — and Still Early
Zscaler's AI Protect suite, announced as a collective offering at the end of January 2026, has reached approximately $100 million in bookings and is being cited by management as one of the healthier growth vectors in the near-term pipeline. The suite has three components: an AI Asset Management tool that identifies all AI applications in use across an organization — including those operating in the background of non-AI applications — an AI Guardrails capability that monitors and inspects real-time communication between users and AI models to prevent sensitive data exposure, and an AI Red Teaming product, sourced from the SPLX acquisition completed a few quarters ago, that continuously monitors model behavior for drift over time.
Critically, these products do not require a customer to already be running Zscaler's core ZIA or ZPA products. Rubin described AI Protect as "a side door" into prospects that might not yet be Zscaler customers, giving the go-to-market team an alternative landing motion at a time when new logo acquisition has been a focus area. How quickly this converts to meaningful ARR remains to be seen, but the fact that it can land independently of the core platform is structurally important.
Non-Seat Revenue Now 30% of New ACV — and Climbing
One of the more concrete data points Rubin offered was that approximately 30% of new ACV in the most recent quarter came from non-seat-based, metered pricing — up from prior periods. This shift matters because the entire agentic and AI security opportunity is inherently consumption-driven. Token volume, traffic, and asset counts are the natural billing units for these products, and Zscaler is already demonstrating that its revenue mix is moving in that direction ahead of the agentic monetization coming online. The Z-Flex contracting vehicle, which crossed $1 billion in bookings this quarter on the back of $480 million in TCV closed, is part of the same structural shift — giving large customers pre-negotiated fixed pricing across the full portfolio with flexibility to flex into new products as they launch.
Fiscal 2027 Guide: Two Specific Headwinds, Not a Business Problem
Rubin was notably transparent about why the early fiscal 2027 outlook disappointed the market. Two senior sales leaders who reported directly to Chief Revenue Officer Mike Rich are departing — one took a role at an AI pre-IPO startup after nearly a decade at Zscaler, and another is also leaving. One replacement will be an internal promotion; the other will likely be an outside hire. Rubin acknowledged the cautious framing directly: "These are not the only directs to Mike, but he doesn't have a large set of directs as well. So just in the interest of being prudent, we took a cautious approach."
The second factor is the pace of uptake for the integrated SecOps offering, which is being migrated from the legacy Red Canary product into Zscaler's native data fabric. More detail on this product is expected at the company's Zenith Live user conference in Las Vegas the following week. Rubin said the uncertainty is not about whether the product has value — it is about "what the pace of that uptake looks like going into next year." Investors should treat the 2027 guide as a conservative baseline with identifiable upside drivers, not as evidence of underlying demand deterioration.
Capital Allocation: Disciplined Tuck-Ins, No Land Grabs
With over $3.5 billion in cash on the balance sheet, Rubin was asked whether a larger strategic acquisition could be on the table. The answer was essentially no, at least for now. "We have really focused on teams and technology that we feel are highly complementary to the Zscaler platform," he said. "We have not sought to expand into adjacencies or buy populations of customers." The Symmetry deal fits this framework precisely — a small team with specific technical capability that slots into the agentic exchange architecture. Investors looking for a transformative M&A moment in the near to medium term are likely to be disappointed, but the capital provides meaningful optionality if the competitive landscape shifts.
The setup into Zenith Live is unusually rich given the number of product threads — agentic exchange, integrated SecOps, AI Protect monetization, and the Glasswing and DayBreak frontier model partnerships — that are expected to see more formal disclosure in the days following the Baird conference. Rubin offered limited detail on Glasswing beyond confirming Zscaler has had access to frontier models and has been running them against its own environments. "You'll hear more from us going forward," he said — which, for now, is the most investors are going to get.