Supabase has introduced AI inference capabilities in its Edge Functions, now powered by Ollama. This update includes embedding models and large language models (LLMs) like Mistral and llama2, directly within the edge runtime, eliminating cold-boot issues. The new features are designed to enhance embedding generation speed, reduce costs, and support seamless AI inference without cold starts. Supabase's Edge Functions are also equipped with a GPU-powered sidecar to bolster performance. The integration aims to improve local development experiences and deployment times, maintaining its commitment to open-source solutions.