When Pinecone debuted a vector database targeting data scientists in 2021, it was most likely too early. As the different applications started to become clearer in 2020, the business began advocating for AI-powered semantic search. With the surge of LLMs being more widely comprehended, organizations are beginning to recognize the benefits of vector databases to a greater degree.
Investors seem to concur; the business recently unveiled a $100 million Series B provision at an estimated value of $750 million. This is quite remarkable in this conservative investing climate. The organization is quickly expanding and investors saw it as a great opportunity to become a competitive force in the market, according to Edo Liberty, the CEO and founder of Pinecone.
Liberty asserted to TechCrunch that they are distinctively the innovators of the vector database and still hold the superiority in the industry as it is now well-established with multiple contenders. It is easy to back the category leader who already has an established spot.
The initial entrance onto the market has allowed Liberty to grow their customer base from only a few last year to 1,500 at present. Liberty claims the rate of expansion mirrors that of a consumer product rather than a difficult to implement database. Organizations of all sizes, including Shopify, Gong and Zapier from the tech industry, have expressed interest.
He remarked that we were experiencing a consumer-oriented adoption of B2B high-tech elements, something he’d not seen before. To progress this, he noted that it would be costly and difficult to expand the capabilities.
He claims that the increasing curiosity in LLMs is raising attention towards the vector database, though they are two completely different ideas. Though both store massive amounts of information so you may sort through it, the LLM model encodes the infomation, making it less adaptable, whereas the vector databse is constructed for finding associations but it has the advantage of being customizable like a regular database.
He explains that if you use a vector-based database rather than the normal database, it is more malleable, and the task of completing GDPR compliance is simpler. He gives an example of this by saying that it is much easier to delete a record in a database than it is to get rid of wrong information in a model architecture because of the format of the model.
Peter Levine from Andreessen Horowitz believes that the vector database is an important factor in the Artificial Intelligence stack. He believes that putting resources behind Pinecone will help it reach the goals of Edo. He told TechCrunch that they are very supportive of this move.
Levine believes that the vector database and the LLMs could collaborate and provide a dependable source of data, potentially addressing the misinterpretation issue associated with LLMs. He remarked, “The two would sync up to give accurate and enduring outcomes. The LLM could be thought of as a type of program that utilizes the database to push information and produce more precise results.” Liberty considers the database to be the LLM’s permanent memory.
The business has been given $100 million in funding and will be hiring more people soon. At the moment there are about 100 workers, and by the end of the year this could reach as high as 200.
Since the launch of Pinecone a few years ago, the area of vector databases has become more and more competitive, with companies like Qdrant, Zilliz, and Chroma recently obtaining financial backing.
Andreessen Horowitz took the lead in today’s investment, with ICONIQ Growth, Menlo Ventures, and Wing Venture Capital (all of which had previously invested) also taking part. The company has now accumulated a total of $138 million; this is days after its $28 million Series A in 2020 and its $10 million seed investment in 2021.
