Secure and optimize your AI journey with F5 and Google Cloud

According to the F5 2025 State of Application Strategy Report, 96% of organizations are currently deploying AI models. Among them, IT leaders cite model security and compute costs as top concerns. It’s easy to see why: new, AI-specific security risks are constantly emerging, and every interaction with an AI system in the cloud can incur new expenses.  

Driven by these factors, enterprises are looking to simultaneously enhance the security of their AI applications, create cost savings, and optimize performance—especially as distributed application architectures that amplify complexity and further increase scale become commonplace.

F5 and Google Cloud are working together to solve these challenges and help remove roadblocks from your AI journey. Here’s how our technologies combine to equip you with a powerful solution for optimizing and securing AI apps.  

Evolving threats such as prompt injection, model poisoning, and data leakage are reshaping the risk landscape—and posing novel challenges for security and application teams. Traditional security frameworks, such as defense-in-depth approaches, are increasingly inadequate against these AI-driven vulnerabilities.

To help protect against emerging risks, F5 and Google Cloud are collaborating to operationalize the Secure AI Framework (SAIF) across the entire AI application stack. In doing so, we’re enabling framework-aligned security that goes beyond traditional approaches.

For example, one of the most critical aspects of AI risk mitigation is securing the user prompts given to models and the outputs generated in response. Using F5 AI Gateway, your team can inspect prompts and responses in real time, and implement adaptive controls to achieve the faster feedback loops emphasized by SAIF. F5 AI Gateway employs dynamic filtering that learns from attack patterns alongside automated threat response that adapts to the latest prompt injection techniques. 

F5 AI Gateway is part of the F5 Application Delivery and Security Platform, a comprehensive solution built for today’s hybrid, multicloud, AI-driven landscape. F5 ADSP equips Google Cloud users with complete security and delivery capabilities for AI apps, consolidated in a single offering to simplify management for IT and security teams. In addition to F5 AI Gateway, F5 ADSP includes capabilities such as API security, AI runtime security, and protected AI data delivery to help safeguard your organization throughout the AI lifecycle.

F5 AI Gateway can also be used with Google Cloud Sensitive Data Protection to automatically detect and redact personally identifiable information or other confidential data from prompts and responses, providing robust, automated data loss prevention for your AI applications.

With AI solutions deployed at scale to both employees and customers, many organizations are experiencing higher-than-expected expenses and unforeseen overruns. In fact, 72% of IT and financial leaders report that their cloud spend on generative AI is becoming unmanageable. 

Integrating F5 AI Gateway with Google Cloud AI infrastructure allows organizations to take advantage of high-performance hardware such as Google Cloud TPUs or NVIDIA GPUs while maintaining control over costs and performance. F5 AI Gateway provides a range of features that enable precise control and optimization for AI applications, including:

• Intelligent load balancing: Distribute AI requests across multiple model instances based on real-time performance metrics, resource availability, and cost considerations.
• Semantic caching: Dramatically reduce costs and improve response times by intelligently caching semantically similar requests.
• Traffic routing optimization: Direct requests to the most appropriate AI model based on complexity, cost, or compliance requirements. 
• Rate limiting: Prevent resource exhaustion and control costs by establishing usage thresholds for departments, applications, APIs, or individual users.
• Edge AI deployment with F5 Distributed Cloud Services: Run AI workloads closer to data sources or users to reduce latency and bandwidth costs while improving application responsiveness.