At Google Cloud Next ‘25, Palo Alto Networks showcased how to secure AI-powered applications in production using AI Runtime Security, our latest innovation in our software firewalls portfolio. Through expert-led sessions, in-booth theater talks and live demos, we showed how to stop threats in real time across AI applications, models and AI agents running on Google Cloud.
Why AI Runtime Security Is Needed – Now
As organizations embrace AI, the risk landscape is evolving faster than traditional security controls can adapt. AI Runtime Security helps teams defend against:
- Prompt injections and model manipulation
- Malicious URLs and database query abuse
- Sensitive data exfiltration and dataset poisoning
These are not hypothetical attacks. We demonstrated how misuse of AI models can compromise the confidentiality, integrity and trustworthiness of critical business systems.
How It Works: Discover, Protect, Monitor
We demonstrated how AI Runtime Security helps Google Cloud teams discover, protect and monitor GenAI applications in real-time at Google Cloud Next '25.
We walked attendees through the three stages of securing AI in Google Cloud with AI Runtime Security:
- Discover: Understand how AI applications and agents interact within your Google Cloud environment. Map data flows track model behavior and pinpoint where risk exists.
- Protect: Stop zero-day threats in zero time. AI Runtime Security uses inline detection to block prompt injection, misuse, adversarial inputs and exfiltration – at wire speed. Teams can deploy using a centralized network intercept or API-based Detection as Code to embed protection into AI apps directly.
- Monitor: Continuously observe your AI stack—including models, agents and user interactions—for anomalies and policy violations. Stay ahead of new attack techniques and protect innovation at runtime.
By combining network-grade defense with developer-grade flexibility, AI Runtime Security helps teams ship AI faster, reduce manual effort, and safeguard sensitive data and models powering AI innovation.
We showed how AI Runtime Security defends ecosystems through real-time protection and deep visibility on Google Cloud.
In the Spotlight: Home Depot + Google Cloud + Palo Alto Networks
Home Depot and Google Cloud shared how they’re securing AI with AI Runtime Security from Palo Alto Networks to enable speed without compromising trust during the Partner Spotlight.
In a featured Partner Spotlight, Home Depot, Google Cloud and Palo Alto Networks explored how to secure AI from the ground up – sharing real-world lessons on Shadow AI, data governance and inline protection during development.
Home Depot described how integrating AI Runtime Security directly into their Google Cloud environment allowed them to move fast without compromising on resilience, trust or control.
Built for Google Cloud AI Teams
AI Runtime Security meets both security and development teams where they are:
- For network teams: Centralized, inline protection across perimeter and east-west Google Cloud traffic.
- For developers: API-based intercepts for simple integration with AI apps – deployed in minutes, managed as code.
Ready to Secure Your AI Stack in Google Cloud?
Start here if you missed us at Google Cloud Next and want to dive deeper:
- Experience a day in the life of a Cloud Network Security Engineer
Walk through our interactive product tour to see how software firewalls protect modern cloud environments. - See AI Runtime Security in Action
Schedule a personalized 1:1 demo and learn how to stop threats in AI environments – live, inline, and in real time.
Palo Alto Networks is redefining what it means to secure AI at runtime in Google Cloud. We’re delivering the platform, flexibility and precision required to stop tomorrow’s AI threats today.