How DigitalOcean Makes Money

DigitalOcean sells cloud infrastructure on usage-based and monthly pricing. Customers pay for Droplet virtual machines, GPU instances, managed Kubernetes, managed databases, object and block storage, bandwidth, App Platform, Cloudways managed hosting plans, and AI inference (billed by usage or tokens). Most customers sign up self-serve with a credit card, which keeps acquisition costs low, while a direct sales team now targets larger "digital native" and AI-native accounts. In 2026 it began signing multi-year commitments, including its first nine-figure annual contracts, which lifted remaining performance obligations (RPO) to $894 million at June 30, 2026, up from $71 million a year earlier.

DigitalOcean's strategy under Paddy Srinivasan has two tracks. First, move upmarket: its number of $100K+ customers grew 9% in Q2 2026 while their revenue grew 98%, and $1M+ customers generated 23% of revenue. Second, become an inference and agentic AI platform: it combines GPU Droplets (NVIDIA and AMD), serverless inference, the Inference Engine, and agent tooling with its existing compute, storage, and databases. It funds this with operating cash flow, committed data center leases, and capital markets activity such as repurchasing about $472 million of its 2030 convertible notes in 2026.