Databricks confidentially files for U.S. IPO, targeting fall debut, sources say
Enzo Reno ·
Databricks has confidentially filed a draft registration statement with the U.S. Securities and Exchange Commission for an initial public offering, according to people familiar with the matter on Wednesday, June 17, 2026. The San Francisco company aims to list as soon as the fall in New York, the people said, seeking to capitalize on sustained enterprise demand for generative AI and data platforms.
If it proceeds on that timeline, the listing would be the most closely watched software float since 2021, setting valuation and growth benchmarks for AI infrastructure peers and late stage startups. Investors will scrutinize Databricks’ margins and consumption trends against public comps such as Snowflake, MongoDB, and Confluent.
Why this matters
Public markets have reopened for profitable or high growth infrastructure names, but AI exposure remains uneven across software. A successful Databricks debut would test whether investors will pay a premium for “lakehouse” platforms that bundle data engineering, machine learning, and model serving on top of hyperscale clouds.
Databricks was valued at $43 billion in a $500 million financing in September 2023 led by T. Rowe Price, with participation from Nvidia and Capital One, the company said at the time. Databricks reported more than $1.5 billion in annual recurring revenue in 2023 and over 10,000 customers, including hundreds with over $1 million in ARR, according to prior company disclosures.
The company has expanded its AI footprint with acquisitions and model releases. It bought MosaicML for about $1.3 billion in 2023 to bring model training in house, acquired Arcion in 2023 for change data capture, and purchased Tabular in 2024 to deepen support for Apache Iceberg. In March 2024 it introduced DBRX, an open large language model positioned for enterprise tasks, and integrated inference and retrieval features across its platform.
Databricks runs natively on Amazon Web Services, Microsoft Azure, and Google Cloud. It partners with Nvidia for accelerated training and inference, and competes with Snowflake, which reported $2.8 billion in product revenue for fiscal 2024, and with Microsoft’s Fabric suite that unifies data warehousing and analytics across Azure.
Under U.S. rules, confidential submissions allow companies to refine financials and risk factors before publishing a full S 1, which typically appears weeks ahead of a roadshow. Banks have not been disclosed, and the filing remains subject to market conditions and SEC review, the people said.
A credible Databricks offering would signal how public investors value AI infrastructure relative to model providers and chipmakers, and it could reset private market pricing for late stage AI startups that rely on cloud and GPU ecosystems controlled by Amazon, Microsoft, Google, and Nvidia.