OpenAI is enhancing its technology stack with the acquisition of Rockset, a vendor recognized for delivering a real-time analytics database that supports intelligent enterprise applications.
While financial details of the acquisition remain undisclosed, OpenAI has confirmed plans to integrate Rockset’s “world-class indexing and querying” capabilities to bolster its retrieval infrastructure across products. Additionally, Rockset’s entire team will join OpenAI as part of this transition.
This acquisition marks OpenAI's second significant acquisition within a year, following its purchase of New York City’s Global Illumination, Inc. The move comes amidst increasing competition in the generative AI landscape; for instance, Anthropic recently launched Claude 3.5 Sonnet, a large language model (LLM) that outperforms OpenAI’s GPT-4o on various benchmarks.
Ilya Sutskever, OpenAI co-founder and former chief scientist, has also initiated his own AI startup, Safe SuperIntelligence.
How Will Rockset Enhance OpenAI?
Founded in 2016, Rockset offers a cloud-based real-time analytics database that empowers developers to create data-intensive applications, particularly for personalization and IT automation, at scale.
Rockset continuously ingests and indexes data from platforms such as Kafka, MongoDB, DynamoDB, and S3, ensuring real-time information availability and seamless querying. It supports sub-second SQL queries on semi-structured data without the need for a predefined schema.
At its core, Rockset utilizes the open-source RocksDB persistent key-value store, developed by Meta (formerly Facebook). It functions as an external secondary index for online transaction processing (OLTP) databases, data lakes, and streaming platforms, enhancing real-time analytic queries while ensuring performance isolation for primary transactional systems. Throughout 2023, Rockset has improved its offerings to better support AI applications.
With OpenAI's acquisition, Rockset's advanced indexing and querying capabilities will now enhance OpenAI's retrieval framework.
Although specific details about the integration are yet to be released, one thing is clear: Rockset will enable OpenAI's products to deliver the most timely and relevant information, significantly improving response speed and accuracy. Imagine an enterprise version of GPT better equipped to address user inquiries with precise data.
“Rockset’s infrastructure empowers companies to transform their data into actionable intelligence. We’re excited to bring these benefits to our customers by integrating Rockset’s capabilities into OpenAI products,” said Brad Lightcap, COO at OpenAI.
Organizations and startups leveraging OpenAI’s models have already been exploring “retrieval-augmented generation,” a technique introduced in a 2020 research paper by Meta, University College London, and New York University. This approach connects generative AI with external knowledge bases, enhancing the model’s capacity to handle specific queries and lowering the chances of incorrect responses—such as providing employees with accurate expense policy information extracted from enterprise documentation.
As for Rockset, the impact on its existing customers after the acquisition remains uncertain. The company has previously collaborated with notable industry players, including Klarna, Meta, Whatnot, and Windward. Responses to media inquiries were not available at the time of writing.
“We’re excited to join OpenAI to empower users, enterprises, and developers to fully leverage their data by integrating powerful retrieval capabilities into AI,” stated Venkat Venkataramani, CEO of Rockset.