Jan 30, 2025 · 1h 6m · mad
Trino, Iceberg and the Battle for the Lakehouse | Justin Borgman, CEO, Starburst
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
On this episode of The MAD Podcast, host Matt Turck interviews Justin Borgman, CEO and co-founder of Starburst, to discuss the evolution of open-source data analytics, the rise of Apache Iceberg and data lakehouse architecture, and the strategic decisions behind building high-scale enterprise data platforms. Borgman shares key insights on rebranding Presto to Trino, navigating hybrid cloud realities through a major OEM partnership with Dell, and balancing open-source community growth with enterprise monetization.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 28.8% of the talking time here. How this is scored →
Matt as informed peer 4.0 Guest teaching 2.5 Guest disagreement 0.2 Matt pushing back 0.5
speaking balance: gold is Matt, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 48:51 Pushing back against Databricks controlling Iceberg
Justin forcefully rejects the premise that Databricks buying Tabular gives them control over Iceberg, arguing the market and Apache Software Foundation maintain strict independence.
Hardest push from Matt ▶ 48:20 Matt questions the logic of the $2B Tabular acquisitionMatt directly challenges the value proposition of spending $2B on Tabular if the buyer cannot control the underlying open-source project.
Biggest teaching moment ▶ 48:51 Justin details open source governance versus corporate buyoutsJustin educates the host on how Apache Software Foundation governance prevents any single vendor from commandeering open format standards.
Matt holds his own ▶ 48:00 Matt reveals internal details of the Tabular acquisition bidding warMatt displays deep domain expertise by citing confidential bidding numbers between Snowflake ($300M to $600M) and Databricks ($2B) alongside strategic motives for Databricks staying private.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Matt as informed peer | Guest teaching | Guest disagreement | Matt pushing back | Why |
|---|---|---|---|---|---|---|
| Elevator Pitch for Starburst and Trino | 3 | 2 | 0 | 0 | Matt sets up the interview with standard framing questions to establish an elevator pitch. Justin cleanly differentiates transactional versus analytical databases in simple terms. | |
| Founder Journey: Hadapt, Yale Research, and Teradata Acquisition | 3 | 2 | 0 | 0 | Matt prompts Justin on his founder history starting from Hadapt at Yale to Teradata. Justin explains how Presto was discovered inside Facebook and the innovator's dilemma faced at Teradata. | |
| Rebranding Presto to Trino and Rebuilding Community Awareness | 2 | 1 | 0 | 0 | Matt asks about the Presto to Trino rebranding story. Justin details how Facebook owned the trademark, forcing the community to rebrand and rebuild GitHub awareness from scratch. | |
| Comparing Data Lakes, Cloud Data Warehouses, and Lakehouses | 6 | 3 | 0 | 1 | Matt demonstrates solid ecosystem expertise by listing vendor categories from his MAD landscape. Justin breaks down columnar file formats and how lakehouse query speeds caught up with warehouses. | |
| Data Federation Philosophy and Navigating Market Overlap | 5 | 2 | 0 | 0 | Matt summarizes Starburst's core philosophy regarding federated data queries. Justin describes how market lines between Databricks, Snowflake, and Starburst are blurring. | |
| Starburst's Core Differentiators and the Icehouse Architecture | 4 | 3 | 0 | 1 | Justin highlights Starburst differentiators including hybrid on-prem support, open Iceberg focus ('Icehouse'), and multi-source join use cases. Matt asks if backing Iceberg early was controversial. | |
| Product Portfolio: Starburst Enterprise, Galaxy, and Dell Partnership | 3 | 2 | 0 | 0 | Matt asks Justin to walk through Starburst's core product lineup. Justin explains their open-core enterprise model and proprietary acceleration features like Warp Speed. | |
| Balancing Open-Source Community and Enterprise Monetization | 4 | 3 | 0 | 1 | Matt probes on the classic open source tension between community contributions and enterprise monetization. Justin outlines his adoption versus conversion framework. | |
| On-Premises Enterprise Reality, Cloud Repatriation, and AI Infrastructure | 3 | 3 | 1 | 0 | Justin highlights why on-premises deployments remain vital for top financial institutions, citing a major bank CEO who insists on keeping four clouds including on-prem. | |
| Dell Data Lakehouse and AI Infrastructure Strategy | 4 | 2 | 0 | 0 | Matt asks about the Dell partnership and real-time streaming capabilities. Justin explains Kafka streaming ingest directly into Iceberg tables. | |
| Core Query Engine, Enterprise Governance, and Table Maintenance | 5 | 2 | 0 | 1 | Matt playfully admits to reading off Starburst's website architecture diagram while asking about governance and table maintenance. Justin elaborates on automated compaction. | |
| Data Applications and Custom Developer Use Cases | 4 | 2 | 0 | 0 | Matt asks about data applications built on top of Starburst compared to embedded BI tools. Justin provides examples like Vectra.ai for cybersecurity analytics. | |
| Re-building Galaxy: The Challenge of Product Re-architecture | 4 | 3 | 0 | 1 | Matt asks what was the hardest product to build in Starburst's history. Justin reveals they scrapped the entire initial version of Galaxy to adopt an Apple-style managed approach, adding a full year of development. | |
| Multi-Product Management and Technical Chassis Strategy | 4 | 2 | 0 | 0 | Matt asks how Starburst manages multi-product PM complexity. Justin uses an automotive VW chassis analogy to explain shared technical components. | |
| Databricks Acquisition of Tabular and Open Source Independence | 7 | 6 | 3 | 3 | Matt demonstrates high industry knowledge dropping specifics on Databricks' $2B Tabular acquisition and bidding war details with Snowflake. Justin delivers a detailed explanation of why Apache foundation governance prevents Databricks from controlling Iceberg. | |
| Open Source Dominance and Historical Enterprise Lessons | 5 | 3 | 0 | 0 | Justin shares historical lessons from Impala vs Teradata/Hadapt, noting open formats always win over proprietary systems long-term. Matt and Justin evaluate the current state of Data Mesh. | |
| Integrating Starburst into Enterprise AI and RAG Workflows | 4 | 3 | 0 | 1 | Discussion turns to enterprise AI, RAG workflows, and go-to-market lessons. Justin admits Starburst learned it was a hardcore enterprise software company rather than a PLG business. | |
| The Dell Partnership: Strategy and OEM Evolution | 4 | 2 | 0 | 0 | Matt asks how the Dell partnership originated given its rarity for startups. Justin details how Dell's product VP approached them directly, leading to a full OEM integration. | |
| Partnership Timeline and Executive Ownership | 3 | 2 | 0 | 0 | Justin outlines executive ownership of strategic partnerships and shares 2025 predictions around production AI and RAG. Matt closes the interview. |
Statements from this episode (27)
Justin Borgman: Starburst created Trino, used by Netflix, Airbnb, and LinkedIn
“We're the creators of an open source project called Trino, which is a pretty popular project used by a lot of the Big internet companies like Netflix and Airbnb and LinkedIn and so forth.”
Borgman: Facebook ran all data warehousing analytics on Presto
“It was really how Facebook was running all of their data warehousing analytics.”
Borgman: Presto is mostly Facebook internal, while Trino is the mainstream branch
“They started as identical copies, but the code bases have diverged quite a bit. And today Presto is really just used by Facebook. So it's sort of like their own private branch in a way used by a small number of people. And Trino has become the mainstream commu…”
Justin Borgman: Data lakehouses are the future of data architecture
“That performance gap is de minimis at this point, and that really has changed the game, and I think now lake houses are the future.”
Justin Borgman: Teradata offers unique database capabilities no competitor can match
“It's an amazing database. I will still say that now, you know, seven years later, it's an amazing database. There are things that that, that system can do that, That really nobody else on the market can do. But, ah, it's proprietary and it's expensive.”
Borgman: Enterprises will never store all data in a single repository
“We do think data lakes are where you're going to want to store as much data as you can, just because the economics will drive that, but you'll never store everything”
Borgman: Databricks and Snowflake expand into adjacent markets to justify valuations
“You have a couple giant players in Databricks and Snowflake that are now looking for adjacent markets to continue to grow their TAM and justify their valuations and drive their revenue into the future. And, ah, and that's leading to more overlaps into boxes th…”
Borgman: Databricks and Snowflake operate strictly as cloud-only data platforms
“Databricks and Snowflake are cloud only, so if you happen to have data on-prem we're pretty much your only bet, and you know, it just so happens that that turns out to be most of the Fortune 500 almost the entirety of the financial Social services sector in pa…”
Borgman: Watching hyper-scale tech adopters predicts broader enterprise infrastructure trends
“You can watch those companies and kind of see where technology is probably going to go because they're the ones running at the most ridiculous scale. These technologies get really tested to the limit that way and can be a good indication of sort of where thing…”
Borgman: Open-source companies should monetize performance and security features
“Things around performance and security are very logical places to kind of draw some lines where you know that enterprise customers value that and you know, that's something that can be monetized.”
Starburst maintains the Apache license for Trino without adding restrictive licensing
“We don't, and that's something that at least philosophically my co-founders have been pretty, I guess, consistent on is a desire to continue to use the Apache license, which gives widespread flexibility and freedom to users of the technology, and so we haven't…”
Borgman: AI emergence may trigger cloud repatriation to on-premises
“There's you know, an interesting case to be made that there may be some repatriation with the AI emergence.”
Borgman: Dell is selling massive quantities of on-prem AI servers
“They're selling, selling a ton of servers right now, AI servers, which are basically, you know, have a lot of GPUs in them, a lot of NVIDIA products in there, and are seeing customers that are trying to get economies of scale by deploying that infrastructure o…”
Borgman: On-prem vs. cloud dichotomy will persist indefinitely in enterprises
“I think this dichotomy is going to exist for as far as I can see.”
Borgman: Data architectures are shifting from batch ETL to Kafka streaming
“And that's an architecture I would say that we're seeing a lot of out there is Kafka as opposed to more traditional batch-oriented ETL.”
Starburst scrapped and rebuilt its Galaxy SaaS platform before launch
“We actually built it twice. And, you know, customers never really saw the first version because at the last minute we said, you know what, we can't deliver the seamless, easy to use, consistent experience that we're going for.”
Starburst was bootstrapped and profitable for its first two years
“We were bootstrapped the first two years, and we were running a nice little profitable business.”
Justin Borgman: Apache Iceberg won the open table format war in 2024
“I feel like the summer of twenty-twenty-four, the world said, okay, it is Iceberg. And that was like VHS, Betamax, You know, decision made.”
Borgman: Databricks will not control or kill Apache Iceberg
“Iceberg has, you know, is, is bigger than Databricks, honestly. And it will continue to be, I think, the ubiquitous format. You know, so if they had hopes of, like, killing it, I don't think that worked. If they had hopes of controlling it, I don't really thin…”
Borgman: Tabular acquisition gives Databricks marketing win, not open-source control
“I think probably the biggest thing that they get out of the acquisition is the ability to tell the market, That they can do iceberg two and not look like they had made a mistake with Dell with Delta. You know, that, that they can, they get sort of a marketing …”
Justin Borgman: Open-source enterprise infrastructure beats proprietary alternatives over time
“I believe, gradually over time, enterprise software technology, at least infrastructure technology, wherever there are two things that are mostly the same, and one is open and one is not, the open is going to win over time.”
Borgman: Data mesh's lasting legacy is the concept of data products
“I think the lasting legacy, ah, of that, though, is this concept of data products, creating these sort of curated data sets from data that can live in, in multiple places, and thinking about them from a product perspective, with a product mindset, which is to …”
Borgman: Starburst is hardcore enterprise software, not a PLG motion
“We're hardcore enterprise software. You know, we're not a PLG motion.”
Borgman: Early software startups should go deep with small boutique SIs
“Rather than going far and wide and signing up 100,000 of SIs, which we did that too along the way I would say go deep with one or two, Probably on the smaller end, just because you'll, you'll, you'll be able to get more attention span and try to make them as g…”
Borgman: Enterprise partnerships must be driven by product or sales
“I think of the partnerships organization, at least in this context, as the facilitators of the relationship, but the real Impetus. The real motivation has to, I think, come from product. It could come from sales.”