The deal is done, quietly
Dremio's own blog now carries a banner stating the company is part of SAP, a closing that landed in September 2026 after being flagged to the market earlier this year with an expected Q3 close. There was no splashy joint press event, which is itself a signal: this is being absorbed as a component acquisition, not launched as a flagship SAP product line. For a company that spent years marketing itself as the independent, engine-agnostic alternative to proprietary lakehouse platforms, that quiet framing is a meaningful shift in identity.
What SAP actually gets is concrete: a distributed SQL query engine with Apache Iceberg support, a caching layer Dremio calls reflections that speeds up repeat queries, and lineage running back through Dremio's MapR and Apache Drill heritage. Folded into SAP's existing data product portfolio, it gives SAP a lakehouse analytics story it did not previously own outright, competing more directly with Databricks, Snowflake, and the open-engine vendors that built their pitch around exactly the independence Dremio is now losing. It also gives SAP customers already running SAP's core applications a more direct path to query Iceberg tables without bolting on a third-party engine.
Open format, proprietary parent
The entire premise of the open lakehouse movement is that Apache Iceberg tables decouple your data from any single vendor's engine. That premise still technically holds after this deal: Iceberg tables written under Dremio remain readable by any Iceberg-compatible engine, including SAP's competitors. But data portability was never the whole pitch customers bought when they chose Dremio. They bought a roadmap, a support relationship, and a company whose commercial interest was making its engine work well with everyone else's stack, not steering customers toward a single vendor's broader suite.
That commercial interest is what changed the moment the deal closed. Starburst, a direct competitor selling its own Iceberg query engine, put the risk plainly: SAP is one of the largest proprietary software companies on earth, with every financial incentive to move customers onto its own stack over time. Discount the source for competitive motive, but do not discount the underlying mechanism. Acquirers routinely deprioritize a subsidiary's cross-platform roadmap in favor of deeper integration with the parent's core suite, and SAP has no commercial reason to keep funding Dremio's neutrality.
The technical claims worth verifying yourself
Starburst also published its own benchmarking claims alongside the acquisition news, arguing Dremio's engine struggled at concurrency levels of just two or more simultaneous queries on large datasets, a limitation it frames as disqualifying for AI agent workloads that fire parallel requests. It further argues Dremio's semantic layer decorates data with descriptions but does not enforce business rules the way its own Context Layer does. Treat both claims as a competitor's marketing, not settled fact, since Starburst has every reason to publish numbers that favor its own architecture.
That skepticism cuts both ways, though. If you are currently running Dremio, or evaluating it against Starburst, Trino, or a cloud-native alternative, the acquisition is your excuse to run your own concurrency and governance tests rather than trusting either vendor's numbers. Ask specifically how the engine performs when five or ten agents query the same large table simultaneously, since that is the workload pattern agentic AI is about to put on every lakehouse in production, regardless of which vendor built it.
A consolidating market, not an isolated deal
This closing lands in the middle of a broader wave of lakehouse consolidation, as large platform vendors acquire the independent query engines and catalogs that grew up around Apache Iceberg's rise over the past two years. The pattern is consistent: a hyperscaler or enterprise suite vendor decides it is cheaper to buy proven Iceberg tooling than build it internally, and a startup that marketed itself on neutrality becomes a feature inside someone else's platform within a budget cycle or two. Customers who built their architecture around picking best-of-breed, vendor-neutral components are watching that neutrality erode one acquisition at a time, often faster than their own procurement cycles can respond to it.
The lesson is not that open formats were a bad bet. Iceberg tables are still more portable than proprietary formats, and that portability is real value that has already paid off for customers migrating between engines. The lesson is that portability at the storage layer does not guarantee independence at the query, catalog, or semantic layer, and CTOs need a distinct governance answer for each of those layers rather than assuming one open standard covers the whole stack end to end.
What to do if you are on Dremio today
If your organization runs production workloads on Dremio, this closing is the trigger to have an explicit conversation with your account team about roadmap commitments, support SLAs, and pricing protection for the next contract renewal, not to wait until those terms change unilaterally. Ask directly whether SAP plans to keep Dremio sold as a standalone, multi-cloud, multi-suite product, or whether it becomes primarily an on-ramp into SAP's broader data and analytics stack. Get the answer in writing before your next renewal cycle, not after.
If you are still evaluating lakehouse query engines, add vendor independence risk as a scored criterion alongside performance and cost, and weight it by the vendor's balance sheet and strategic position, since a well-funded independent is a safer long-term bet than a recently acquired one regardless of today's benchmarks. The Iceberg ecosystem is maturing fast enough that this will not be the last acquisition of its kind, and the CTOs who plan for that pattern now, building exit ramps into their contracts before they need them, will spend far less time renegotiating under pressure later than the ones who assumed today's vendor map was permanent.
