Over the past year, there has been tremendous growth in the development and applications of advanced artificial intelligence technologies. Now, a startup called Exa is tackling an important challenge that this rising trend presents – how to ensure AI systems have easy and efficient access to the vast amounts of online information they will increasingly need to draw from.
Exa, founded by Harvard alumni Will Bryk and Jeff Wang, has developed a search engine platform designed specifically with the needs of AI models in mind. Where traditional search engines like Google work primarily based on keyword matching and ranking web pages, Exa trains machine learning models to understand relationships between online content at a deeper level. By predicting logical connections and sequences between different information sources on the web, such as one page linking to another, Exa aims to surface the most relevant results for AI queries rather than just the most popular or well-optimized pages.
After launching their API version around a year ago, Exa says it has gained significant traction with thousands of developers now using it. Notable customers include Databricks, who leverages Exa to locate large datasets useful for their own AI training efforts. With AI continuing to spread into new industries and applications, the demand for reliable and tailored information access for these systems will only increase.
Reflecting the promise of their novel approach, Exa has now raised a $17 million Series A round led by venture capital firm Lightspeed. Also participating were Nvidia, the world’s leading AI chipmaker, and renowned startup accelerator Y Combinator. This new funding will help Exa further develop their technology while expanding their commercial efforts to more fully realize their vision of a search platform built specifically for artificial intelligence. With AI poised to transform numerous sectors of the economy, Exa aims to play a key foundational role by ensuring these transformative technologies have optimized access to the broad range of online knowledge they will increasingly rely on.
