A10 Networks is set to launch its A10 AI Gateway in the fourth quarter of 2026, introducing a centralized platform designed to help enterprises manage, secure and control the growing use of artificial intelligence across their organizations.
The company said the A10 AI Gateway will serve as an intelligent control plane for managing AI agents, applications and large language models (LLMs), addressing challenges that have emerged as businesses increasingly integrate generative AI and AI-powered agents into their operations.
As organizations adopt models from multiple AI providers, IT teams face the increasingly complex task of tracking which models are being used, monitoring their costs and ensuring that appropriate governance and security policies are enforced.
A10 Networks said its new platform is designed to consolidate these functions into a single management environment, providing organizations with greater visibility over AI usage while helping optimize performance, latency, security and spending.
At the core of the A10 AI Gateway is an intelligent routing system that evaluates incoming requests and determines which AI model should handle them based on factors such as the content and complexity of the request and the organization’s policies.
Simpler tasks, for example, can be automatically directed to less expensive models, while requests requiring more advanced reasoning can be routed to higher-performance models. The approach is intended to help enterprises balance AI performance with operating costs without requiring users to manually select models for individual tasks.
The gateway also supports identity-based access controls through integration with existing local directory systems. Organizations can establish different AI access and routing policies for individual users or groups, allowing departments such as engineering and finance to use different models or operate under separate usage rules.
Cost management is another major component of the platform. A10 said the gateway provides real-time, per-request cost tracking in dollars, as well as token budget management for individual models and teams. Administrators can establish hard spending limits, soft-limit alerts and rate controls based on tokens or requests per minute.
These capabilities are designed to provide enterprises with a clearer picture of how much their AI workloads cost and where resources are being consumed.
The A10 AI Gateway will also integrate with the company’s broader AI security portfolio, including TrojAI, A10 AI Firewall and ThreatX. According to A10, the combined technologies are intended to provide security throughout the AI lifecycle, from pre-deployment testing and runtime protection to controls at the network and application programming interface (API) layers.
The platform can be deployed as standalone software or with integrated hardware and is designed to operate within a customer’s own infrastructure. Supported environments include on-premises systems, private clouds and air-gapped networks.
A10 said the deployment options are particularly suited to organizations that need to retain direct control over proprietary AI models, sensitive corporate information and governance policies instead of relying entirely on cloud-hosted AI infrastructure.
“Enterprises are being asked to move fast on AI and stay in control at the same time. AI use is spreading faster than teams can discover, govern, or secure it, driving up cost and leaving no single place to set the rules,” A10 Networks President and CEO Dhrupad Trivedi said.
“The A10 AI Gateway gives organizations one intelligent control plane to see how AI is being used, to govern who can access which models, and match every request to the right model at the right cost,” he added.
Trivedi said the company’s integrated portfolio is aimed at giving customers a “complete, sovereign way to put AI to work at scale,” allowing enterprises to expand their use of artificial intelligence without sacrificing organizational control.
The A10 AI Gateway is scheduled to become available in the fourth quarter of 2026 as enterprises continue to look for ways to manage the rapid proliferation of AI models, agents and applications across their operations.