Databricks agent stats and IPO chatter
A market write‑up cited that only 19% of organizations have deployed AI agents, while platform-centric usage examples claim agents were creating the majority of databases on some systems and multi-agent workflows grew rapidly. The broader coverage also repeats speculation about a potential Databricks 2026 IPO with a reported $134 billion valuation and $4.4 billion in revenue. Those numbers combine platform adoption signals with public‑market interest in data infrastructure firms. (saastr.com) (tech-insider.org)
Databricks’ own 2026 agent report says only 19% of organizations have put artificial intelligence agents into production, even as agent activity is rising fast on its platform. (databricks.com) (saastr.com) The report draws on data from more than 20,000 organizations worldwide, including over 60% of the Fortune 500, and Databricks said companies are shifting from chatbots to systems that can plan, call tools and take actions. Databricks also said companies using governance tools moved 12 times more artificial intelligence projects into production, while users of evaluation tools moved nearly six times more systems into production. (databricks.com) On the infrastructure side, outside coverage of the report said artificial intelligence agents now create 80% of new databases on Neon and 97% of database branches, while multi-agent workflows grew 327% between June and October 2025. Those figures describe activity inside Databricks-linked systems, not adoption across all companies. (forbes.com) (paperclipped.de) A database is the system that stores records for an application, and a branch is a copy developers use to test changes before touching the live version. Databricks has tied that trend to Lakebase, its serverless Postgres product for applications and artificial intelligence agents. (databricks.com) (bloomberg.com) That split helps explain the mixed headline numbers. Few organizations have deployed agents broadly, but the ones that have are generating heavy usage in narrow tasks such as customer support, market intelligence, regulatory reporting and software workflows. (databricks.com) (siliconangle.com) The public-markets angle comes from Databricks’ financing pace and valuation, not from a filed offering. TechCrunch reported on December 16, 2025 that Databricks raised more than $4 billion at a $134 billion valuation and said the company had more than $4.8 billion in revenue run rate, up 55% from a year earlier. (techcrunch.com) Earlier, Bloomberg reported on June 11, 2025 that Databricks was projected to post $4.3 billion in product revenue in the fiscal year ending in January 2026. The same report said Databricks SQL was expected to reach a $1 billion revenue run rate by that fiscal year end. (bloomberg.com) Databricks has not publicly filed a registration statement for an initial public offering, and recent reporting still frames 2026 as possible rather than scheduled. Chief executive Ali Ghodsi told CNBC in December 2025 that he “wouldn’t rule out” going public in 2026, while saying there was no fixed commitment. (cnbc.com) (economictimes.indiatimes.com) Ghodsi had struck a more cautious tone a year earlier. TechCrunch reported on December 17, 2024 that he said it was “dumb” to go public in 2024 and that Databricks wanted more stability before listing. (techcrunch.com) So the story is two separate clocks running at once: enterprise agent deployment is still limited, while Databricks’ platform metrics, fundraising and revenue scale are feeding fresh speculation about a 2026 debut that the company still has not formally set. (databricks.com) (techcrunch.com)
