Advertisement
UK markets closed
  • FTSE 100

    8,203.93
    -37.33 (-0.45%)
     
  • FTSE 250

    20,786.65
    +176.31 (+0.86%)
     
  • AIM

    774.39
    +4.97 (+0.65%)
     
  • GBP/EUR

    1.1819
    +0.0021 (+0.18%)
     
  • GBP/USD

    1.2813
    +0.0052 (+0.41%)
     
  • Bitcoin GBP

    44,966.34
    +623.41 (+1.41%)
     
  • CMC Crypto 200

    1,191.65
    -17.05 (-1.41%)
     
  • S&P 500

    5,567.19
    +30.17 (+0.54%)
     
  • DOW

    39,375.87
    +67.87 (+0.17%)
     
  • CRUDE OIL

    83.44
    -0.44 (-0.52%)
     
  • GOLD FUTURES

    2,399.80
    +30.40 (+1.28%)
     
  • NIKKEI 225

    40,912.37
    -1.28 (-0.00%)
     
  • HANG SENG

    17,799.61
    -228.67 (-1.27%)
     
  • DAX

    18,475.45
    +24.97 (+0.14%)
     
  • CAC 40

    7,675.62
    -20.16 (-0.26%)
     

Databricks CEO Ali Ghodsi on Snowflake rivalry and the ‘why’ behind Databricks’ latest billion-dollar deal

David Paul Morris/Bloomberg—Getty Images

Here we are again, I tell Databricks CEO Ali Ghodsi as we hop on a Zoom on a sunny Tuesday.

It really does feel like déjà vu. Databricks has been on a headline-making acquisition streak over the last year, and this week the company announced another big-ticket deal with its plan to buy open source-focused data manager Tabular.

Though we don’t know the exact contours of the deal, Databricks confirmed it's north of $1 billion, making it the company’s second billion-dollar deal within 12 months. Databricks made a splash in late June 2023, buying MosaicML for $1.3 billion. And the deals have rolled out quickly since: In October, Databricks bought Arcion, a Databricks Ventures-backed startup, in a deal valued at $100 million. In January, Term Sheet had the exclusive on Databricks’ buyout of AI-focused data platform Einblick, and in March the company bought Lilac AI.

The Tabular deal is noteworthy for a few reasons. First, for its size: Databricks is now officially in the “we do billion-dollar deals repeatedly” club, which is pretty extraordinary for a VC-backed startup. Second, the deal reflects the growing rivalry in a high-stakes business data turf war between Databricks and Snowflake.

ADVERTISEMENT

The name of the latest chapter in this rivalry: Iceberg.

Iceberg is an open-source layer that helps store and organize massive amounts of data in “data lakes” (for the uninitiated, a data lake is a low-cost but unstructured way of storing raw data). Iceberg was started in 2017 by Ryan Blue and Dan Weeks, who with Jason Reid, cofounded Tabular in 2021. Tabular would go on to be backed by Andreessen Horowitz, Altimeter Capital, and Zetta Venture Partners, raising $37 million through Series B, according to Crunchbase.

A source familiar with the matter told me that Snowflake also tried to buy Tabular, and ended up locked in a bidding war with Databricks. (Snowflake declined comment for this story). I ask Ghodsi for his take on how Tabular fits into the competition with Snowflake:

“What I will say is this: I think that this lakehouse approach, where you own your own data is obviously super disruptive,” he said. “So, yeah, I think it’s a challenge to them, but frankly there are so many others out there… To be honest with you, I think they all see it as threatening to them, because they’re going to lose revenues. Snowflake itself has said that they’re going to lose revenue because of this lakehouse open source format.”

Snowflake seemed to acknowledge the Iceberg challenge in its Q4 earnings call, when its CFO warned investors of “revenue headwinds” due to, among other things, “the expectation that some of our customers will leverage Iceberg tables for their storage.”

But that was back in February, and a lot has changed in three months, including Snowflake’s CEO. The company is now led by Sridhar Ramaswamy, and on Monday Snowflake announced its Polaris Catalog—a bet on an enhanced open source strategy that’s directly tied to(you guessed it) Iceberg. (If you want to read more about Snowflake, my colleague Sharon Goldman just interviewed Ramaswamy and wrote a piece about his vision—including on M&A).

And lest you think any of this isn’t inextricably linked to AI, never fear—it is.

“The lakehouse, this Tabular [deal], these kinds of things actually enable you to store other types of data, and that’s important for AI,” said Ghodsi. “It’s the data ingredient of AI. It’s not the algorithm or the ChatGPT thing. It’s the data. And I think people undervalue data in this space, for AI.”

But understanding the timeline on all this is tricky, Ghodsi said.

“In the short-term, everybody's overestimating the impact of these disruptions,” he told Term Sheet. “In the long run, everybody's underestimating the impact of these disruptions. I think that’s what’s happened… The crossroads for us is, how well can we really embrace generative AI in the fullness of the platform and really make it super, super easy for anyone to use these platforms?... That will completely determine our future.”

See you tomorrow,

Allie Garfinkle
Twitter:
@agarfinks
Email: alexandra.garfinkle@fortune.com
Submit a deal for the Term Sheet newsletter here.

Joe Abrams curated the deals section of today's newsletter.

Correction: The email version of this newsletter had a typo in Ghodsi's name. We regret the error.

This story was originally featured on Fortune.com