Okta's latest earnings reveal an important shift in how the market is thinking about artificial intelligence.
AI agents do not only need intelligence.
They need identities.
That idea helped push Okta shares more than 19% higher in after-hours trading after the identity-security company reported stronger-than-expected fiscal second-quarter results and improved its outlook.
Okta generated $805 million in Q2 revenue, up 11% year over year, with subscription revenue rising 12% to $793 million.
Remaining performance obligations increased 17% to $4.858 billion, while current RPO increased 14% to $2.585 billion.
Those numbers showed improving underlying demand.
But CEO Todd McKinnon's comments on AI agents may explain why investors see a larger opportunity developing.
Every AI Agent Needs Permission to Do Something
Traditional identity software solves a familiar problem.
A company needs to know:
- who an employee is, - whether that employee should be allowed to log in, - which systems they can access, - and what they are authorized to do.
AI agents create essentially the same problem—but at machine scale.
An AI agent may be authorized to query customer records, approve a workflow, generate code or interact with financial systems.
That means enterprises need controls determining what each agent can access and which actions it can perform.
The long-term opportunity is straightforward: every AI agent needs a trusted identity and explicit controls over what it can access and do.
This is a fundamentally different use case from simply securing employee passwords.
Human Identity Is Becoming Machine Identity
The identity-security market has already expanded beyond workforce authentication.
Companies now manage:
Human identities — employees and customers.
Machine identities — servers, applications, APIs and automated processes.
And increasingly:
AI-agent identities — autonomous software capable of reasoning and taking actions.
If autonomous agents become common inside large organizations, the number of identities requiring governance could grow far faster than the number of employees.
That is the long-term opportunity investors are starting to consider.
The Existing Business Is Also Improving
The AI narrative would be less convincing if Okta's core financial performance were deteriorating.
Instead, several metrics strengthened.
Q2 GAAP operating income increased to $107 million, or 13% of revenue, from $41 million a year earlier.
Free cash flow reached $227 million, representing a 28% margin.
The company also raised its fiscal 2027 revenue outlook to approximately $3.216 billion to $3.226 billion, representing growth of 10% to 11%.
Full-year non-GAAP free cash flow is expected to reach $910 million to $930 million.
That combination—moderate revenue growth plus stronger margins and cash generation—is particularly attractive for a software company that previously prioritized growth much more aggressively.
Identity May Become a Bottleneck for Enterprise AI
Generative AI models are improving rapidly.
But enterprises cannot simply give an autonomous model unrestricted access to internal systems.
A capable AI agent with excessive permissions can create risks similar to a compromised privileged employee account.
Potential problems include:
- leaking sensitive data, - modifying records incorrectly, - accessing unauthorized systems, - executing harmful actions, - or being manipulated through prompt injection.
That means AI deployment may increasingly depend on identity and governance infrastructure.
The more capable agents become, the more important those controls become.
This is why identity could emerge as one of the less obvious infrastructure winners of the AI cycle.
Okta Still Faces Real Challenges
The AI-agent opportunity is early.
There is no guarantee that identity spending associated with agents becomes large enough to materially accelerate Okta's overall revenue growth.
Microsoft also possesses a major competitive advantage through Entra and its broader enterprise ecosystem.
Cybersecurity vendors such as CrowdStrike are expanding further into identity protection, while cloud platforms may offer more native AI-agent controls.
Okta therefore needs to prove that being an independent identity layer remains strategically valuable as the market evolves.
Its current 10%–11% revenue growth rate is healthy, but it is far below the hypergrowth phase investors once associated with the company.
Why the Earnings Reaction Makes Sense
Okta's nearly 20% post-earnings move is not simply a reaction to an $805 million revenue quarter.
Investors are reconsidering the size of the identity market itself.
The first era of identity security was about employees logging into software.
The next may involve humans, applications, machines and millions of autonomous agents interacting simultaneously.
If that future develops, identity becomes less like a login feature and more like a core control plane for AI-enabled enterprises.
Okta has not proven it will win that market.
But its latest results suggest the company's core business is strengthening at exactly the moment a new identity category is emerging.