Pinecone, the vector database startup founded by Edo Liberty, has made significant strides in helping businesses enhance large language models (LLMs) with their own data. Recently, the company underwent a complete product rearchitecture, resulting in the launch of Pinecone Serverless, a solution that eliminates the need for customers to manage their deployments and scale them. After a successful beta period, Pinecone Serverless is now generally available.
Edo Liberty, the founder of Pinecone, acknowledged that their early customers have transitioned from experimenting with generative AI to wanting to launch their own AI products. This shift presented challenges for enterprises as they had to navigate the complexities of building new applications and putting them into production.
Liberty explained that Pinecone’s more than 5,000 customers expressed a need for a dedicated and specialized tool that excels in vector search, RAG (retrieval-augmented generation), knowledge extraction, and context generation for language models. Essentially, they required scalability, high performance, and cost-effectiveness to develop their AI products successfully.
To meet these demands, Pinecone invested significant time in preparing its product for production deployments while also significantly reducing costs. Through a rearchitecture process, Pinecone created a multi-tenant service that separates storage and compute. As a result, customers using Pinecone Serverless can reduce their costs by up to 50 times. The company achieves this by charging customers only for the CPU time they consume, with the capacity orchestrated in the backend.
Pinecone’s approach of running everything as a service enables them to accurately charge customers for their usage, offering a level of cost optimization that is rare and challenging to achieve in the industry.
During the public preview phase, Pinecone received valuable feedback from customers and incorporated additional features into its offering. One such feature is Private Endpoints, which allows enterprises to establish a direct connection to their virtual private clouds on Amazon through AWS PrivateLink. This connection ensures that data remains within the organization’s governance and compliance frameworks, as it doesn’t traverse the public internet.
Notable companies that have already embraced Pinecone Serverless include Gong, Help Scout, New Relic, Notion, TaskUS, and You.com. Notion, in particular, praises Pinecone for enabling their AI feature to provide instant answers to millions of users from billions of documents. By adopting Pinecone’s latest architecture, Notion has achieved a 60% cost reduction, furthering their mission to make software toolmaking ubiquitous.
Pinecone’s focus on optimizing performance, scalability, and cost-efficiency positions them as a valuable partner for enterprises seeking to leverage AI and language models effectively. With the availability of Pinecone Serverless and the addition of features like Private Endpoints, the company continues to address the evolving needs of its customers, helping them unlock the full potential of AI technology in their respective industries.
