The rapid rise of enterprise AI is forcing IT teams to rethink how they connect and protect their systems. As autonomous AI agents multiply across networks, a new kind of digital gatekeeper is emerging. These AI agent gateways manage the chaos, control costs and secure sensitive data.
"The game has changed. It is not just about APIs anymore," Ashwini Vasanth, product lead for Nutanix Enterprise AI, told The Forecast.
"With AI, the economics have shifted to tokens as the main currency. IT teams must now track and govern these tokens to maintain control over their infrastructure in addition to managing what agents actually do."
What is an AI Agent Gateway?
An AI agent gateway is an inline control plane that sits between every AI agent and its target systems. It routes traffic, enforces access policies, tracks token consumption and records every interaction for audit and compliance purposes. Unlike a traditional API gateway, which manages data flow between microservices, an AI agent gateway governs the behavior of autonomous agents that can initiate actions, chain decisions, and interact with multiple LLMs databases, and business tools in real time.
Vasanth is part of the team behind the Nutanix AI Agent Gateway, a centralized control layer within the Nutanix Enterprise AI (NAI) platform, that gives IT teams a grip on everything AI touches inside the enterprise.
She describes an AI gateway vs. API gateway. API gateways emerged around 2014 to manage traffic flowing between microservices and enforce data policies. A decade later, AI agent gateways are redefining those capabilities to track agentic AI applications across different IT infrastructures, data sets and large language models (LLMs).
How are AI Agent Gateways Used?
From all indications, AI agents are no longer experimental. They are already growing exponentially. Gartner expects 40% of enterprise applications will embed task-specific AI agents by 2026, up from less than 5% in 2025. Furthermore, Cisco's State of AI Security 2026 report found that 83% of organizations planned to deploy agentic capabilities, but only 29% felt prepared to secure them. IDC's Future Enterprise Resiliency and Spending survey from January 2026 found that AI and agent security and governance now account for 16.7% of average worldwide planned AI investment.
"AI agents are multiplying, and the governance layer needs to keep pace," Vasanth said. "Positioning a gateway between every AI agent and its target systems allows IT teams to observe, control and optimize token usage."
Vasanth explained that every agent builder deploys a gateway alongside its agent. This addresses the inherent challenges of reaching multiple providers via bespoke APIs.
“Every admin is trying to track down the users and agents consuming AI within the organization,” she said. “With a centralized agent gateway, developers innovate on a shared gateway that is monitored by IT.”
IT organizations will need to bolster the infrastructure that runs and governs their digital capabilities, explained Vasanth. They will need agent gateways that are scalable and modular, given unpredictable consumption patterns and rapid technology iteration.
The Nutanix Agent Gateway, for example, allows AI developers to use a variety of different model providers without worrying about the underlying infrastructure. Vasanath explained that it doesn’t require specialized hardware.
"The agent gateway does not have that constraint,” she said. “It can be used on-premises or in the cloud. IT teams can test it or keep it running in AWS or another cloud service."
Tuhina Goel is the go-to-market lead for Nutanix Enterprise AI (NAI). She explained that the Agent Gateway in NAI also supports emerging connectivity standards, including the Model Context Protocol (MCP). This allows agents to interact with business tools such as GitHub and Stripe through a standardized interface, eliminating the need for bespoke integrations with each provider.
Goel compared the gateway to a combination of a traffic cop, customs officer and flight recorder for the enterprise AI stack, all compressed into an invisible control plane.
"When an agent calls a tool, queries a database or communicates with another agent, that call passes through the gateway first," Goel said.
Every interaction is recorded for auditing against system governance policies. This helps organizations comply with emerging regulations, such as the Digital Operational Resilience Act (DORA), which requires strict audit trails and access controls.
"Agent Gateway gives organizations the policy enforcement, access controls and audit trails needed to deploy AI responsibly,” Goel said. “That reduces the risk of compliance failures, data breaches and audit gaps that carry significant regulatory consequences.”
Goel said this is especially critical for regulated industries such as federal, healthcare and financial services.
Why Enterprise IT Needs an AI Agent Gateway
Industry insiders see a new norm in which AI innovation is proliferating new capabilities and increasing complexity.
"I cannot think of a timeframe where there has just been this combination of scale of what is unknown as well as rapid evolution in innovation happening all at the same time," Scott Sinclair, practice director at Omdia, said.
"It is unprecedented. As you start to adopt new technologies or integrate new initiatives, it is all about how do we take this unknown and turn it into known so we can ensure that we are putting the right level of security and controls on top of it, and doing the right things we can to best protect our business."
Sinclair said agents are evolving rapidly inside IT organizations, but each one has a different level of risk appetite to how they want agents to work and how quickly they're able to adopt them. IT pros have to learn a lot of new things nowadays.
How AI Agent Gateways Control Token Costs and Shadow AI
Tokenomics is the design and study of the economic rules, supply metrics and incentive structures that govern a cryptocurrency ecosystem. It’s growing more widely used with the rise of enterprise AI. Tokenmaxxing quickly became a cost concern, as some employees were using AI excessively.
Axios reported in late May that an unnamed enterprise client racked up a $500 million bill on Anthropic's Claude AI in a single month after failing to set usage limits or spending caps for its employees. The report stated that an AI consultant said one of their clients recently spent half a billion dollars in a single month after failing to put usage limits on Claude licenses for employees.
An agent gateway helps IT teams protect their systems. For example, the popular coding assistant Claude Code is tied to Anthropic's model by default. However, employees can use Claude, but IT can route it to a different LLM. For the IT team governing resources, the LLM provider becomes a configurable policy decision rather than a hardcoded dependency.
"You can use Claude Code with the Nutanix Agent Gateway and it doesn’t matter what model it plugs into," Vasanth said. “IT can control the token usage and also ensure access to enterprise-approved models.”
The developers' experience does not change because the entire governance control is abstracted, according to Goel.
"AI workloads can be maintained in a private, controlled and sovereign manner without letting users realize there is a change in experience," Goel said.
She said this approach helps IT teams limit shadow AI, which is often funded by other business units but introduces vulnerabilities that eventually burden IT systems teams.
AI Agent Gateways: The New Foundation for Enterprise AI Governance
Deciding to deploy an AI agent gateway comes down to intertwined dynamics, including credential sprawl and visibility, explained Vasanth and Goel. Without a gateway, the systems admin must create a key for each provider and each user or agent. This huge sprawl of keys creates management challenges. Without a central gateway for observing usage, organizations monitoring AI adoption lack a shared view to evaluate their AI adoption.
A gateway connects agents to systems by defining and auditing the boundaries for where they can run and what they can touch.
"Agent gateways will become an integral part of enterprise AI strategies," Vasanth said.
"It brings visibility and control. That visibility helps organizations evaluate adoption targets and identify token optimization options.”
Related:
- Introducing Nutanix Agent Gateway: Unified Governance and Cost Control for Agentic AI
- One Gateway. Three Estates. Zero Compromises.
- Effective AI Governance for Organizations
- ECI 2026 Report Shows Strain Between AI Innovation and IT Governance
- How to Develop Top-Down Enterprise AI Strategies
- ModelOps CTO On Strong Governance Is Essential for Agentic AI
- Shadow IT Surges as Employees Deploy Unsanctioned AI Tools and Agents
Editor's note: Learn how Nutanix Enterprise AI technology enables IT teams to connect, govern, and run AI across hybrid environments, with a centralized Agent Gateway and inferencing for agents and models at scale.
Ken Kaplan is Editor in Chief for The Forecast by Nutanix. Find him on X @kenekaplan and LinkedIn.
© 2026 Nutanix, Inc. All rights reserved. For additional information and important legal disclaimers, please go here.
Related Articles
