Fastly is at the forefront of helping enterprises adapt to the operational demands of AI, serving as a critical infrastructure layer where performance and security challenges are addressed head on. Fastly delivers one million requests per second* on behalf of the open source projects we support.

Because our edge platform is central to how these workloads are delivered and secured, we have a direct view into the technical friction points of the Agentic Web. To address these hurdles, Fastly has joined the Agentic AI Foundation (AAIF) as a member to ensure that the protocols governing agentic interactions are built for the scale and speed of real-world production environments.

Our primary technical focus centers on MCP and the secure execution of agent sub-tasks. Having already released a Fastly MCP Server to connect LLMs to our edge tools, we are further committing to contribute to the evolution of the protocol to ensure it remains optimized for high-frequency, low-latency edge operations. We help our customers get the most out of MCP every day, and we’re excited to advocate for them and the broader agentic web with this membership.

Integrating autonomous agents into your production environment benefits from a shift of manual workflows to machine-assisted execution that doesn’t compromise on security or performance. Our participation in the AAIF is focused on ensuring your agentic architectures benefit from the sub-millisecond execution and data sovereignty required for enterprise-grade deployments. 

This commitment is a natural extension of our Fast Forward program, through which we provide in-kind infrastructure and support to the critical open source projects that power the web. By reducing the thought latency of every agentic action, from automated cache invalidation to dynamic WAF adjustments, we are building the high performance, standardized foundation you need to scale AI-driven operations safely. 

Build using our Fastly MCP server today, and tell us what you need from the next developments for shaping standards that actually work in production. If you're figuring out how to optimize your infrastructure for the transition to agents, reach out to ensure you’re getting the most out of the performance and security benefits of the edge.

*Based on annualized data from 10/20/24 to 01/17/25