Fly.io introduces GPU capabilities to its cloud service platform for AI applications, including support for Nvidia A100 GPUs.
The service allows for full-stack applications to run closer to users with easy setup, and includes automatic scaling features for cost efficiency.
Users can deploy applications in global datacentres to ensure low-latency performance, with same IP and TLS certificates across regions.
Pricing for the new GPU offerings is provided with a teaser of upcoming hardware, including discounts for reserved instances.
The update emphasizes the company's commitment to enhancing AI workloads and invites users to share their experiences on their forum.

Fly.io specializes in transforming containers into micro-VMs that can be rapidly deployed across 30+ regions worldwide, promising near-instantaneous setup from source code. Their platform supports full Linux micro-VMs, providing enterprise-level security through hardware isolation and a memory-safe stack. They offer globally distributed databases that are easy to manage, without the complexity, and feature fast nVME storage with automated backups. Fly.io facilitates Postgres clustering with efficient leader election and member management. The service is developer-friendly, with support for popular frameworks, Dockerfile generation, and compatibility with various databases, streamlining the deployment process.

Fly.io deploys containers as micro-VMs globally for fast app delivery.

Fly.io Launches GPU Support for AI Edge Workloads

Recent Product News