Neon has faced challenges in providing autoscaling for a serverless database and dealing with 'cold starts', but has made significant progress in these areas. The company is also enhancing its AI capabilities by storing vector embeddings and has developed its own vector extension, pg_embedding, to improve accuracy. CEO Nikita Shamgunov stated that Neon's strength lies in being PostgreSQL, which allows users to store the majority of their data in PostgreSQL and perform vector searches without needing a separate database.
Key takeaways:
- Neon, a serverless PostgreSQL database company, has raised $46 million in a series B funding round, bringing its total funding to $104 million.
- The company's service provides the open source PostgreSQL relational database as a serverless cloud service, eliminating the need for developers to maintain servers.
- Neon has overcome challenges such as autoscaling for a serverless database and 'cold starts', reducing its cold start time to sub-200 milliseconds.
- Neon is enhancing its AI capabilities with its own vector extension called pg_embedding, which provides vector search and uses modern algorithms for improved accuracy.