Cloudflare, a global connectivity cloud company, has announced new integrations that will allow developers to easily and affordably deploy Artificial Intelligence (AI) applications on its platform. In a partnership with Hugging Face, they offer a serverless, one-click, AI application deployment solution.
This new feature is a gamechanger in the fast-paced AI development world. Users no longer need to manage infrastructure or pay for unused compute capacity, making it more affordable and convenient for organisations to build production-ready AI applications. Furthermore, Workers AI, which is the only at-scale platform for running AI inference at the edge, has attained GA (generally available) status with GPUs rolled out in over 150 cities worldwide, indicating Cloudflare's commitment to the next wave of internet development.
Matthew Prince, CEO, and co-founder of Cloudflare, explained the challenges that the AI sector has faced and the need for such a solution. "The recent generative AI boom has companies across industries investing massive amounts of time and money into AI. Some of it will work, but the real challenge of AI is that the demo is easy, but putting it into production is incredibly hard," he said. "We can solve this by abstracting away the cost and complexity of building AI-powered apps."
Workers AI is hailed as a significant step in the democratisation of AI. It is equipped to efficiently handle production workloads and offers a promising solution that combines speed, security, performance, and observability. Furthermore, the platform is continuously being expanded to support fine-tuned model weights, which will help organisations develop more specialized domain-specific applications.
The new Python support on Workers AI, which is now in open beta, aligns with Cloudflare's goal of widening its support for Python packages. Python is the second most popular programming language after JavaScript, and its seamless integration into the Cloudflare platform is viewed as a major advancement.
The partnership with Hugging Face is another notable highlight providing a one-click deployment solution for AI models. Cloudflare's global serverless inference platform, optimised for 14 curated Hugging Face models, enables faster access to a diverse range of models and instantaneous deployment across Cloudflare’s global network of GPUs.
Julien Chaumond, co-founder and chief technology officer at Hugging Face, expressed his excitement over the collaboration, stating, "Offering the most popular open models with a serverless API, powered by a global fleet of GPUs, is an amazing proposition for the Hugging Face community, and I can’t wait to see what they build with it.”
Cloudflare's AI-focused developments have already been welcomed and adopted by various companies, including Talkmap, ChainFuse.com, Discourse.org, Simmer, and Audioflare, highlighting the platform's potential for fostering a new era of innovative AI applications.