AWS re:Invent 2025-From idea to instant production readiness:building AI agents on Cloudflare-AIM119

Summary of AWS re:Invent 2025 Presentation: "From idea to instant production readiness: building AI agents on Cloudflare"

The Challenge of Building AI Applications

• Developing AI-powered applications is a complex and time-consuming process, taking an average of 29 weeks and costing around $1 million per project.
• A significant portion of this time and cost is spent on provisioning servers, managing infrastructure, and choosing the right deployment regions, rather than on actual application development.
• This slows down the development process and prevents engineers from focusing on building the core functionality.

Cloudflare's Developer Platform

• Cloudflare has built a developer platform that provides a comprehensive set of building blocks for modern application development, including compute, data, media, and AI capabilities.
• Cloudflare's global network of over 335 data centers, owned and operated by the company, allows developers to deploy their applications close to end-users, improving performance and reducing latency.
• The Cloudflare developer platform is designed to be easy to use, with a focus on modern architectures and developer experiences. Developers can deploy applications with a single command, without having to manage infrastructure.

Cloudflare's Approach to AI Workloads

• Cloudflare made a strategic decision to focus on AI inference rather than training, recognizing that the majority of AI workloads are inference-based.
• Cloudflare has installed GPUs in over 210 cities across its global network, allowing developers to run AI inference workloads close to the end-users, reducing latency and improving performance.
• Cloudflare's pay-as-you-go pricing model for GPU usage and its approach to only charging for actual compute time (not waiting time) help to optimize costs for AI workloads.

Building Agentic AI Applications on Cloudflare

• Cloudflare is now focusing on enabling the development of "agentic AI" applications, where AI agents are given specific goals and tasks to complete autonomously.
• Cloudflare provides a range of building blocks, such as CRM agents, voice agents, and observability tools, to help developers build and deploy these AI-powered applications.
• The Cloudflare platform's global reach, pay-as-you-go pricing, and ease of use make it an attractive option for developers looking to build and deploy AI agents quickly and cost-effectively.

Real-World Example: PayPal's MCP Server on Cloudflare

• PayPal built an MCP (Merchant Commerce Platform) server on the Cloudflare platform, allowing small business owners to easily integrate PayPal services into their operations.
• Using Cloudflare, PayPal was able to build and deploy this MCP server in just 3 days, a significant improvement over the typical 29-week development timeline.
• The MCP server allows small business owners to automate tasks like invoice generation, tax calculation, and accounts receivable/payable management, providing a cost-saving and efficient solution.

Conclusion

Cloudflare's developer platform, with its focus on modern architectures, global reach, and AI capabilities, is well-positioned to help organizations quickly and cost-effectively build and deploy AI-powered applications and agents. The platform's ease of use and pay-as-you-go pricing model make it an attractive option for developers looking to accelerate their AI initiatives.