DigitalOcean

DigitalOcean provides cloud computing and AI infrastructure for developers, startups and small-to-medium-sized businesses. Its platform spans Droplet virtual machines, managed Kubernetes, databases, storage and application hosting. The Inference Engine lets teams run AI applications through serverless, asynchronous batch or dedicated GPU deployments, using catalog models or bringing their own. Developers can test multimodal models, evaluate outputs and track usage, latency and spending alongside their application infrastructure.

Founded in 2012, DigitalOcean began with easy-to-deploy virtual machines; Ben Uretsky was a co-founder, and Paddy Srinivasan is its current CEO. Its AI engineering extends that managed-infrastructure approach to model selection. Inference Router, available in public preview, matches conversation context to natural-language task descriptions, then ranks eligible models using cost or latency data. Cache-aware routing preserves model bindings across related requests and weighs the additional input-token cost of switching models.

DigitalOcean acquired Katanemo Labs in 2026, bringing in the team behind Plano, the open-source proxy powering Inference Router. Plano’s architecture separates intent detection from model ranking and translates between provider APIs. Teams can use DigitalOcean’s managed router or self-host Plano with their own provider keys. DigitalOcean’s inference service uses pay-as-you-go pricing, connecting model access and deployment choices with its broader cloud platform.