At the recent Next '26 event, significant updates to Cloud Run were announced, showcasing its evolution as a robust platform for developers. With a doubling of external active developers and applications last year, Cloud Run is poised for even greater growth in 2025.
Key new features include:
- Full-stack app development: Developers can now build and deploy full-stack applications using Google AI Studio, Firestore, and user authentication.
- Gemini Enterprise Agent Platform: This new platform allows for the creation, scaling, and optimization of AI agents integrated with Cloud Run.
- Remote MCP server: A fully managed remote Model Context Protocol server has been launched to simplify deployments for developers and agents.
- NVIDIA GPU support: High-performance inference is now possible with NVIDIA RTX PRO 6000 Blackwell GPUs on Cloud Run.
Empowering Developers
Cloud Run aims to democratize software development, enabling users to transition from prototypes to deployed applications in mere seconds. The new features in Google AI Studio facilitate the development of full-stack applications with ease.
Agent Integration
AI agents require reliable compute environments to function effectively. The integration with the Gemini Enterprise Agent Platform allows for seamless transitions from experimental setups to production-grade systems.
Cloud Run Instances
Developers can now create individual Cloud Run instances, enhancing control over infrastructure management. This feature is particularly useful for hosting long-running background agents.
Sandboxing Capabilities
A new sandbox tool will allow agents to execute code in a secure, isolated environment, enhancing safety during code execution.
Automatic Scaling and SSH Support
Cloud Run automatically scales to meet demand, making it ideal for handling traffic spikes. Additionally, secure shell (SSH) access to running containers is now available for advanced troubleshooting.
AI Workloads and GPU Support
Cloud Run supports high-end AI workloads with the introduction of NVIDIA RTX PRO 6000 Blackwell GPUs, allowing for efficient model serving and training.
With these advancements, Cloud Run continues to position itself as a leading choice for developers looking to harness the power of serverless computing.
