Singapore – Cloudflare has announced an expansion of its Agent Cloud platform, introducing a range of tools aimed at supporting the development, deployment and scaling of AI agents across its global network. The update reflects a broader shift in the technology landscape, where the focus is moving from access to AI models towards the infrastructure required to run them efficiently, securely and at scale.
As adoption of AI increases across Southeast Asia and beyond, developers are encountering new challenges associated with compute intensity, cost management and operational complexity.
AI agents, which are designed to execute multi-step tasks and operate autonomously, require significantly more dynamic and scalable environments than traditional software applications. Conventional cloud architectures, largely developed for earlier generations of mobile and web applications, are often not optimised for these emerging workloads.
“The way people build software is fundamentally changing. We are entering a world where agents are the ones writing and executing code,” Matthew Prince, Co-founder and CEO of Cloudflare, stated.
“But agents need a home that is secure by default, scales to millions instantly, and persists across long-running tasks. We’ve spent nine years building the foundation for this with Cloudflare Workers. Today, we are making Cloudflare the definitive platform for the agentic web.”
The expanded platform introduces an isolate-based runtime designed to execute AI-generated code within secure, sandboxed environments. This approach is intended to improve efficiency and reduce costs compared with container-based systems, while enabling rapid execution of short-lived tasks such as API calls, data transformations and chained operations. The runtime is designed to scale to high levels of concurrency without the need for pre-initialisation.
In addition, the platform incorporates a new storage capability compatible with Git-based workflows, enabling large-scale repository creation and management. This feature is intended to support the growing volume of code generated by AI agents, providing persistent storage that can be accessed using standard development tools.
“Cloud agents are quickly becoming a foundational building block for how work gets done, and with Cloudflare, we’re making it dramatically easier for developers to deploy, production-ready agents powered by GPT-5.4 and Codex to run real enterprise workloads,” Rohan Varma, Product, Codex, at OpenAI, shared.
The update also includes the general availability of persistent sandbox environments, which provide isolated Linux-based systems for more complex tasks. These environments allow agents to replicate developer workflows, including cloning repositories, installing dependencies, executing builds and iterating on outputs within a consistent execution context.
The platform further integrates a broader range of AI models following Cloudflare’s acquisition of Replicate. Developers can select from both proprietary and open-source models through a unified interface, switching between providers without significant operational overhead.
