Simon Eskildsen, founder of Turbopuffer, revealed on X on May 21 that the company’s Annual Recurring Revenue (ARR) surpassed $100 million in March this year. It took just 19 months to reach this milestone from hitting $1 million in ARR, and Turbopuffer is currently profitable despite having raised less than $1 million in external funding. Its client roster includes leading AI and enterprise software firms such as Cursor, Anthropic, Notion, Cognition, Harvey, Bridgewater, Ramp, Linear, Legora, Superhuman, Atlassian, and Granola. Turbopuffer is a vector database provider headquartered in Ottawa, Canada; its architecture relies on Object Storage as the underlying layer, which reduces storage costs by roughly 90% compared to conventional vector databases in large-scale scenarios.
Such rapid growth is quite rare in the AI infrastructure sector — achieving over $100 million in annual revenue while staying profitable with less than $1 million in funding underscores the surging demand for vector retrieval in Retrieval-Augmented Generation (RAG) and semantic search applications. To date, Turbopuffer has secured large-scale production deployments from numerous major AI applications, earning increasingly positive feedback within developer communities.