Cloudflare said on March 24, 2026 that Dynamic Workers let developers execute AI-generated code inside secure, lightweight isolates and that the approach is 100 times faster than traditional containers. Cloudflare’s blog says the feature is now in open beta for paid Workers users and can block direct outbound internet access with <code>globalOutbound: null</code>.
Cloudflare said on March 24, 2026 that it is working with Arm to deploy the Arm AGI CPU across its global network. Arm's newsroom says the chip is the company's first production silicon product and is aimed at AI data center workloads such as accelerator management, control planes, and API hosting.
Cloudflare said on X on March 19 that Kimi K2.5 is now available on Workers AI. The launch pairs a frontier open-source model with platform features aimed at lowering latency and cost for agent workloads.
Cloudflare said on March 20, 2026 that Kimi K2.5 is now available on Workers AI so developers can run agents end-to-end on its platform. The linked Cloudflare blog says the model ships with a 256K context window, multi-turn tool calling, vision, and structured outputs, and that one internal agent workload cut costs by 77% after the switch.
Cloudflare said on March 20, 2026 that Kimi K2.5 was available on Workers AI so developers could build end-to-end agents on Cloudflare’s platform. Its launch post says the model brings a 256k context window, multi-turn tool calling, vision inputs, and structured outputs, while an internal security-review agent processing 7B tokens per day cut costs by 77% after the switch.
Cloudflare said on March 19, 2026 that Workers AI now supports Moonshot AI's Kimi K2.5. The company is using the model to argue that a unified agent platform can offer both strong tool use and much lower production cost.
On March 11, 2026, Cloudflare announced the general availability of AI Security for Apps. It also made AI endpoint discovery free for Free, Pro, and Business customers, while adding custom-topics detection and integrations involving IBM and Wiz.
Cloudflare said on March 11, 2026 that it now returns RFC 9457-compliant Markdown and JSON error payloads to AI agents instead of heavyweight HTML pages. In a same-day blog post, the company said the change cuts token usage by more than 98% on a live 1015 rate-limit response and turns error handling into machine-readable control flow.
Cloudflare used X on March 9, 2026 to launch an open beta of its Web and API Vulnerability Scanner for API Shield customers. The company says the product uses AI-built API call graphs to actively test BOLA and other logic flaws that standard defensive tools often miss.
Cloudflare says Cloudflare One now links data security controls from endpoints to AI prompts. The update adds browser RDP clipboard controls, richer SaaS operation logging, on-device DLP, and Microsoft 365 Copilot scanning through API CASB.
A high-ranking Hacker News thread highlights Cloudflare’s vinext launch: an AI-assisted, Vite-based reimplementation of the Next.js API surface. The performance claims are notable, but the project is still explicitly experimental.