Databricks Asks for a Declaration That It Does Not Infringe Nom Nom's ETL Patent.

The complaint

The case is Databricks, Inc. v. Nom Nom Data Inc., No. 3:26-cv-09634 (N.D. Cal., San Francisco Division), a 12 page complaint for declaratory judgment of non-infringement filed September 4, 2026, ECF No. 1, with the patent attached. Databricks, a Delaware corporation based in San Francisco, sues under 28 U.S.C. 2201 and 2202. The complaint alleges, on information and belief, that Nom Nom Data Inc. is a Delaware corporation based in Austin, Texas, that Nom Nom AI Inc. is a British Columbia corporation, and that Nom Nom AI Inc. is the assignee of the patent; it says the USPTO lists Nom Nom AI Inc. as the current assignee.

The patent

U.S. Patent 12,298,995, Systems, Methods, and Computer-Readable Media for Managing an Extract, Transform, and Load Process, issued May 13, 2025; the complaint says the application was filed February 21, 2024. Claim 1 covers a computing system with processors and memory whose instructions cause the processors to receive ETL event data from the execution of a set of ETL tasks, identify from that data a first task, corresponding to one or more SQL operations, that limits the performance of the process, identify a limiting resource (memory, storage, processor bandwidth, or network bandwidth), split the first task into two or more subtasks, and automatically modify the task's instructions to dynamically schedule the subtasks as part of the process to improve its performance. The complaint says Nom Nom's charts have mapped claims 1 to 14 and 19 to Databricks products and services.

The pre-suit communications

The complaint describes an October 21, 2025 email to Databricks' chief executive pitching an acquisition discussion with an attached patent analysis naming Databricks features, a November 5 follow-up, and a December 9, 2025 claim chart mapping claims 1 to 10 and 19 to Databricks products and services. It says Nom Nom identified outside counsel on January 15 and July 6, 2026 and stated that law firms had agreed to represent it on contingency; sent a corrected claim chart on February 17 on claims 1 to 5, 10, 12 to 14, and 19, with an 18 company licensing analysis alleging industry exposure of up to $38.7 billion a year; and on July 15 sent an updated chart mapping claims 1 to 6 to the integrated Databricks Data Intelligence Platform. It says that on an August 19 call Nom Nom presented licensing options and said the numbers would be large.

The non-infringement theory

Databricks alleges that its Data Intelligence Platform and services do not process ETL event data to identify an ETL task limiting the performance of the process, because they have no way of determining whether a task is limiting performance, and do not split an executing ETL task into subtasks or modify its task instructions to schedule new subtasks. It says the other independent claims, 10 and 19, recite substantively identical limitations.

Jurisdiction and the relief sought

Databricks pleads personal jurisdiction from Nom Nom's communications directed to California, citing Trimble Inc. v. PerDiemCo LLC, 997 F.3d 1147 (Fed. Cir. 2021), and, for Nom Nom AI Inc., Federal Rule of Civil Procedure 4(k)(2). It asks for a declaration that it does not and has not infringed any claim of the patent, directly or indirectly, costs, a finding that the case is exceptional under 35 U.S.C. 285 with attorneys' fees, and other relief. It demands a jury trial.

The docket so far

The court issued an initial case management scheduling order on September 8, 2026. The docket shows a clerk's notice of impending reassignment and a reassignment on September 23, proofs of service filed October 20, and an order on a pro hac vice application on October 22.

What to watch next

Watch for Nom Nom's response to the complaint, including any motion to dismiss for lack of personal jurisdiction or of a case or controversy, and any infringement counterclaim. The case can end at any time, including by settlement or voluntary dismissal.