Cloud computing has long been promoted as a democratizing force in IT, offering scalable resources and flexible pricing for organizations of all sizes. However, the reality is more nuanced: while cloud services enable innovation, they also introduce significant governance and cost challenges. As enterprises accelerate their adoption of artificial intelligence, particularly agentic AI—where autonomous agents interact with models, applications, and data—these challenges multiply. Runaway token usage, underutilized reserved instances, and over-provisioning can erode the financial benefits of cloud computing. Nutanix, a leader in hybrid multicloud infrastructure, has introduced a new solution to address these pain points: Nutanix Agent Gateway.

Announced as part of Nutanix Enterprise AI 2.7, which reached general availability this week, Nutanix Agent Gateway serves as a centralized front door for managing interactions between AI agents, large language models (LLMs), and enterprise tools. The service is designed to provide platform teams and AI developers with unified control over agent activity, access policies, and token consumption across agentic AI deployments. This move comes as organizations transition from experimental AI pilots to production-scale implementations involving hundreds or thousands of autonomous agents.

The need for such a solution stems from the rapid evolution of agentic AI. Unlike traditional AI workflows that involve simple query-response interactions, agentic AI systems operate autonomously, executing complex tasks that require continuous access to models, enterprise applications, and business data. This autonomy introduces new governance risks, including unauthorized access to sensitive tools, difficulty in tracking costs across multiple model providers, and a lack of visibility into agent behavior. Nutanix Agent Gateway aims to mitigate these risks by offering a single, centralised control point.

Centralized Front Door for Agentic AI

At its core, Nutanix Agent Gateway acts as a policy enforcement layer that sits between AI agents and the resources they need to access. It integrates with Nutanix Enterprise AI to secure interactions between agents, models, and business applications, ensuring consistent governance whether organizations use frontier models hosted in the public cloud or self-hosted private models. The service provides key capabilities such as unified observability, audit logs, and a unified API for accessing both external and self-hosted models.

Speaking at a media gathering in London, Nutanix CEO Rajiv Ramaswami emphasized the importance of controlling AI tool usage. "The industry has been evolving so rapidly, so we have focused on our own productivity experiences to drive our own product development. What we see now is widespread adoption of AI in our employee base, but the catch is that tools are getting more expensive, so with tokenisation in mind, we have looked at how to approach optimisation the right way," Ramaswami said. He stressed that organizations need to put controls on tools to define who has policy privileges to use which tools and at what times. "Simple jobs should only be executed on simple models; we need to move away from the free-for-all model that has reigned up to now."

Ramaswami also discussed how engineering teams should manage model execution at the widest level while maintaining granular focus. He called for cutting what he described as "unfettered use" of not just AI, but all cloud services. While noting that small-scale deployments may not require such management, he underscored that Nutanix's enterprise focus makes the solution relevant for large organizations facing cost and governance pressures.

Key Capabilities: Token Observability and MCP Governance

Nutanix Agent Gateway addresses the cost challenges by centralizing token observability across model providers. IT and platform teams can monitor usage, allocate costs, and control AI spending with greater precision. The visibility also helps organizations identify workloads that can be shifted to self-hosted models, reducing reliance on external services and optimizing costs. This is particularly valuable as AI token consumption becomes a major budget line item for enterprises.

The service also introduces governance for Model Context Protocol (MCP) servers, enabling cloud-native developers and operations teams to set granular access controls. This ensures that agents can securely connect to business tools and private data sources without compromising security. Key functions include:

• Unified Observability: Centralize visibility into token usage, MCP server access, and LLM activity across the organization.
• Audit Logs: Record every MCP request with a comprehensive audit trail, supporting compliance and governance requirements.
• Unified API: Access external provider models and self-hosted models through a single API, allowing developers to choose the right model for each use case without rewriting code.

Additionally, Nutanix Agent Gateway provides granular token-based rate limiting, allowing administrators to enforce token quotas and limits centrally. This delivers real-time visibility into token usage across every agent and team, helping prevent runaway costs.

Industry Context and Developer Impact

The launch of Nutanix Agent Gateway reflects broader trends in enterprise AI deployment. According to Sammy Zoghlami, SVP EMEA at Nutanix, "Organizations are rapidly moving from pilot projects to large-scale agentic AI deployments involving hundreds or even thousands of autonomous agents. Without centralised governance, it becomes difficult to control costs, access, and compliance." Zoghlami emphasized that as autonomous agents proliferate, a unified governance framework becomes essential.

From a software development perspective, Nutanix Agent Gateway offers developers both freedom and operational guardrails. By using a unified API, coders can switch between public and self-hosted models to fit the right use case without rewriting code or introducing unnecessary complexity. The centralized front door handles Model Context Protocol access, real-time token tracking, and rate limiting, allowing developers to focus on building production-scale autonomous code, agents, and connectors.

The service also supports consistent governance across hybrid environments, whether organizations rely on public cloud-hosted frontier models or on-premises, self-hosted models. IT teams benefit from unified management of access policies, governance controls, and token consumption across their AI environments. This approach aligns with Nutanix's broader strategy of simplifying hybrid multicloud operations while addressing the unique demands of AI workloads.

Nutanix has a track record of delivering infrastructure solutions that bridge on-premises and cloud environments. The company's Enterprise AI platform, now at version 2.7, provides a foundation for running AI workloads with built-in data services and security. The addition of Agent Gateway extends this platform to explicitly address the governance and cost challenges of agentic AI, a rapidly growing area of enterprise technology.

As AI agents become more sophisticated, the need for centralized control will only increase. By providing granular token observability, policy enforcement, and audit capabilities, Nutanix Agent Gateway positions itself as a critical tool for organizations seeking to scale agentic AI deployments responsibly. The service is now generally available as part of Nutanix Enterprise AI 2.7, offering enterprises a path to bring order to the complex landscape of autonomous agents and AI model consumption.

Source: Computerweekly News