{"slug": "f5-wants-to-keep-your-ai-safe-and-compute-costs-cheap", "title": "F5 wants to keep your AI safe and compute costs cheap", "summary": "F5 Inc. expanded its F5 AI Security Platform with AI Gateway, a control plane that adds token monitoring, model access controls, MCP governance, and AI guardrails to help enterprises manage AI traffic security and costs. The company claims Model Gateway can cut token spending by up to 60% without application changes, and the platform addresses shadow AI by restricting agent access and auditing MCP server usage. F5 Chief Product Officer Kunal Anand said the tool provides a single control point for managing AI across models, clouds, agents, and applications.", "body_md": "F5 expanded its enterprise AI control panel with advanced token monitoring via its mainline AI-centric security suite.\n\nAvailable in F5 AI Security Platform, AI Gateway touts model access controls, [model context protocol (MCP)](https://www.sdxcentral.com/sdx-explainers/what-is-the-model-context-protocol-mcp-the-definitive-guide-to-ais-new-integration-standard/) governance, and AI guardrails, with policies set centrally and enforced across distributed infrastructure. F5 said the approach is intended to give AI platform, security, and finance teams a common view of AI usage across their software estate.\n\nF5's Model Gateway allows users to track token consumption down to the provider, model, team, and user level while allowing organizations to set spending limits and determine which AI models should handle particular requests. The company is also using routing, model tiering, semantic caching, and graphics processing unit (GPU)-aware load balancing to help direct requests toward the most appropriate model and infrastructure.\n\nF5 framed the tool as a response to increasing pressure on businesses to monitor the security implications of AI traffic and the economic side as each request to a commercial or internally hosted model can consume tokens and compute resources, while enterprises may simultaneously be routing traffic across multiple model providers.\n\nThe security vendor claimed that Model Gateway can help users cut token spending by as much as 60% without requiring application changes, although ultimate savings would depend on the customer's workload and deployment.\n\n## Taking on shadow AI\n\nF5’s AI Gateway also addresses the rapid expansion [of shadow AI activity](https://www.sdxcentral.com/analysis/keeping-shadow-ai-from-the-enterprise-end-zone/) via AI agents and the MCP servers used to external tools and data. The MCP Gateway strand lets administrators restrict which resources individual agents can access and creates an audit trail showing which agent accessed a resource, when it did so, and on whose behalf. F5 is also adding an MCP server registry intended to provide developers and administrators with a central catalogue of approved servers and tools.\n\nSecurity controls are handled through F5 AI Guardrails, which inspect prompts and model responses for potentially sensitive information and malicious activity. F5 said the system can redact sensitive data before it is submitted to a model, block prompt-injection, jailbreak attempts, and reject requests entirely when they cannot be evaluated.\n\n\"We're watching enterprises race to deploy AI while struggling to control it. Every AI request carries economic, security, and governance implications, yet most organizations are relying on fragmented tools that address only part of the problem. The result is rising costs, increased risk, and operational complexity,” F5 Chief Product Officer Kunal Anand said. “F5 AI Gateway, integrated into the F5 AI Security Platform, provides a single control point for managing AI across models, clouds, agents, and applications. We believe every enterprise will need an intelligent control layer for AI. F5 is building that foundation, helping customers accelerate innovation while maintaining visibility, security, and control.\"\n\nF5 AI Gateway can be deployed in software-as-a-service (SaaS), hybrid SaaS, and hybrid multicloud environments. F5 also plans to add air-gapped deployment support, targeting regulated organizations and sovereign environments that cannot allow AI traffic or associated control data to leave isolated infrastructure.", "url": "https://wpnews.pro/news/f5-wants-to-keep-your-ai-safe-and-compute-costs-cheap", "canonical_source": "https://www.sdxcentral.com/news/f5-wants-to-keep-your-ai-safe-and-compute-costs-cheap/", "published_at": "2026-08-19 12:41:12+00:00", "updated_at": "2026-08-19 13:10:51.332410+00:00", "lang": "en", "topics": ["ai-infrastructure", "ai-policy", "ai-tools"], "entities": ["F5 Inc.", "F5 AI Security Platform", "AI Gateway", "Model Gateway", "MCP Gateway", "F5 AI Guardrails", "Kunal Anand"], "alternates": {"html": "https://wpnews.pro/news/f5-wants-to-keep-your-ai-safe-and-compute-costs-cheap", "markdown": "https://wpnews.pro/news/f5-wants-to-keep-your-ai-safe-and-compute-costs-cheap.md", "text": "https://wpnews.pro/news/f5-wants-to-keep-your-ai-safe-and-compute-costs-cheap.txt", "jsonld": "https://wpnews.pro/news/f5-wants-to-keep-your-ai-safe-and-compute-costs-cheap.jsonld"}}