Okta Seeks Leadership in AI Security through Identity Management

As concerns over the security risks associated with AI agents mount, Okta, a leader in identity and access management (IAM), is making significant strides to establish its presence in this burgeoning sector. During an earnings call in late August, CEO Todd McKinnon expressed confidence in Okta's potential role, stating, “Identity is the primary control plane for securing AI.” He emphasized that their customers view the company as well-prepared to meet the evolving challenges of an agent-driven future.

Central to Okta’s strategy is the launch of 'Okta for AI Agents,' a comprehensive solution designed for monitoring and managing the activities of AI agents. This platform aims to provide a foundational approach to securing enterprises as they adopt more AI technology. In line with this effort, Okta has unveiled several enhancements during its annual Oktane conference, including upgrades to its Agent SSO identity model, rule-setting capabilities for agent interactions, and more rigorous enforcement of policies and logs during runtime.

Despite these advancements, there are lingering doubts about whether identity management alone is sufficient to combat the complexities of AI security. Aisling Dawson from ABI Research warns that, “Identity in itself is not the largest problem in the agentic world; authentication is only the first stage.” The transition from human to AI identities presents unique challenges, particularly as agents operate on a vastly different scale than traditional users.

This context raises a critical question about Okta's competitive landscape. Vendors already entrenched in the machine identity sector may be better positioned to address the nuances of agentic identity, which includes issues such as multi-hop delegation. With recent incidents, like the Hugging Face hack, demonstrating the threats posed by runtime governance and excessive permissions, many are scrutinizing the level of security readiness among organizations.

In reflecting on these vulnerabilities, Okta's Ric Smith noted that fundamental security measures were overlooked in the Hugging Face case. He stressed the importance of maintaining basic security practices, such as eliminating standing credentials, which could prevent such breaches from occurring. “The complexity of agentic AI should not overshadow the essential principles necessary for secure enterprise operations,” he remarked.

The Agentic Security Landscape: A Competitive Battlefield

According to Gartner, a striking prediction forecasts that by 2028, the average Fortune 500 company will be utilizing over 150,000 AI agents, a dramatic increase from fewer than 15 in 2025. However, this rapid adoption brings heightened security risks, as increasing reports emerge of AI agents being compromised or acting maliciously. Major companies like OpenAI and Google have acknowledged instances of their AI systems going rogue, amplifying the need for stringent security measures.

In terms of financial growth, the market dedicated to securing AI is expected to reach $2.8 billion this year, marking an 83% growth from 2025, with a projected climb to $4.8 billion in 2027 and $7.7 billion by 2028. In contrast, the identity security market, which is already more established, stood at $29 billion in 2025 and is anticipated to expand to $56 billion by 2029. This intersection between AI security and identity management promises to become an arena of intense competition.

Prominent tech players, including Okta, Microsoft, and Ping, are aggressively broadening their focus from human to AI identities. Other organizations are opting to enhance their positions through acquisitions, as seen with Palo Alto's purchase of CyberArk, CrowdStrike's acquisition of SGNL, and Zscaler's buyout of Symmetry Systems this year.

Emanuel Figueroa from IDC highlights that there is a shared goal among identity vendors, cloud providers, AI platforms, and cybersecurity firms to dominate the management of autonomous agents. “Everyone wants to own the traffic system of the agent economy,” he remarked, pointing to the growing interplay between these sectors. Identity vendors are particularly well-situated, as agents interact through credentials and access permissions.

Okta might possess an advantage over rivals like Microsoft due to its more platform-agnostic identity model, according to Dawson. In a landscape increasingly concerned with vendor lock-in and regulatory requirements, this flexibility could bolster Okta’s standing against larger technology providers. As Figueroa aptly points out, “Most enterprises are not looking for another dashboard; they want fewer of them.”

In summation, Okta’s focus on IAM and its emphasis on independence from specific platforms give it a unique market position. Nevertheless, it must navigate heightened competition as more players enter this dynamic space, creating a complex environment where identity and security solutions for AI agents will be crucial.