Your Company Will Soon Have 150,000 AI Agents

The average large company will have 150,000 AI agents by 2028. A new platform wants to help you manage them all.

Enterprises are facing a new kind of infrastructure challenge: AI agent sprawl. According to VentureBeat, companies are deploying autonomous AI agents much faster than they are building systems to manage them. This creates a significant governance gap. Research from Gartner highlights the staggering scale of this shift, estimating that the average Fortune 500 company will use more than 150,000 AI agents by 2028, a massive jump from fewer than 15 in 2025. Despite this explosive growth, only 13% of organizations feel they have adequate AI governance in place today. To address this emerging crisis, a new startup named xpander is launching a platform designed to give enterprises a unified control layer for their growing fleets of AI agents.

At its core, xpander offers what it calls a “control and context layer.” This is essentially a centralized command center for a company's diverse and distributed AI agents. Instead of managing each agent in a silo, the platform provides a single place to set policies, monitor behavior, and manage permissions. This system works by abstracting the underlying complexity of different AI models and development frameworks. Think of it like a Kubernetes for AI agents; it orchestrates and governs them regardless of where they run or how they were built. This layer also manages “context,” ensuring that agents have secure and appropriate access to the company data and tools they need to perform their tasks without overstepping their boundaries, which is critical for both performance and security.

This problem directly affects CTOs, IT leaders, and security teams who are on the front lines of managing this new complexity. Without a central governance system, a company can quickly lose track of the hundreds or thousands of agents operating within its environment. This chaos introduces serious risks, including data leakage from poorly configured agents, compliance violations if agents handle sensitive information improperly, and security vulnerabilities if agents are compromised. For developers, the lack of a standardized framework can slow down innovation, as each new agent requires custom security and data access solutions. A control plane promises to mitigate these risks by enforcing consistent rules and providing clear visibility into all agent activity.

The emergence of AI agent sprawl signals the creation of a new, essential category in the AI infrastructure market. Just as the explosion of virtual machines and containers led to the rise of cloud management and orchestration platforms, the proliferation of AI agents is creating a demand for sophisticated governance tools. For businesses, this isn't just a technical problem; it's a strategic one. Companies that fail to manage their AI agents effectively risk seeing their AI investments turn into liabilities. The practical takeaway for any organization scaling its AI efforts is to prioritize a governance strategy now. Establishing a control plane early can provide the foundation needed to innovate safely and transform AI from a source of potential chaos into a well-managed, strategic asset.

Looking ahead, the focus will likely shift from simply building individual agents to managing entire ecosystems of them. The success of platforms like xpander will depend on their ability to remain model-agnostic and integrate seamlessly with existing enterprise security and IT systems, such as identity management and observability tools. We are witnessing the birth of a new discipline that could be called “AgentOps,” focused on the lifecycle management, security, and orchestration of autonomous AI systems. How companies choose to own and operate this control layer will become a key competitive differentiator in the years to come, separating the leaders from those overwhelmed by the complexity they created.

Unmanaged AI agents create significant security, compliance, and operational risks for IT and security teams. A central control plane is becoming essential for any company scaling its use of AI to avoid chaos and maintain control over its technology stack.

The rise of AI agent sprawl is creating a new, critical category of AI infrastructure focused on governance and management. This presents a major opportunity for new platforms and a strategic imperative for enterprises to adopt governance tools to de-risk their AI investments.