Databricks outbids Snowflake and Confluent on Tabular, the open-core company behind Apache Iceberg that we covered before. The competition in the modern data stack is still red hot. Quite a surprise to see Tabular exiting so quickly, previously I thought they are of great potential, albeit with monetization challenges. This is a super aggressive move from Databricks because clearly $SNOW and $CFLT are better buyers with clearer synergies. This is more of a defensive strategy for Databricks because it already owns Delta Lake. But as we've analyzed last Feb, it is not going well vs. Iceberg who is destined to become the industry standard, and Delta Universal Format doesn't help that much. The listed goal is to bring in better interoperability between Iceberg and Delta Lake. This is weird because if everything goes well, they should centralize on Delta Lake only, and if it is SNOW or CFLT buying it, they can further supercharge the Iceberg capabilities rather than maintaining two routes. $1bn+ price tag is quite expensive for this purpose. So it seems like Databricks is seeing a huge threat and pressure from Iceberg because it is a true open and flexible standard vs. Delta Lake. Also, this doesn't look like a sound capital allocation strategy, also MosaicML at the benefit of the hindsight isn't of that high ROI. This makes me worry about Databricks' operating efficiency and capital allocation strategy, especially for future public stockholders. But I do think Databricks is better off by buying Tabular, it is just that this is quite expensive as a defensive move. Would be interesting to track what's their next move. Maybe part of the reason why SNOW didn't chase Tabular deal that aggressively is because they have already announced Polaris Catalog and making it open-source. That is killing Tabular's open-core business, and also threatening Databricks' Unity Catalog that is not open-source, and was leveraged as a hook to keep Delta Lake open in surface but close for enterprise usage. For this round 2 of open-source play, it looks like SNOW has played it better and targeting Databricks' pain point. But it is continuously evolving so Databricks will possibly fight back too.