NGraph

Databricks: company structure analysis (Arche)

D

Databricks

Earth 30·USAUSA, 2013
84
/100

Society

8/10

Customer

10/10

Self

10/10

9 above the excellence line

Survival 47Excellence 75

The founders of Apache Spark built a $190B privately-held giant.

12 criteria

50/60

+

3 axes

28/30

+

Bonus

+6

=

NCG

84/100

Bonus: [Impact+Capital] $190B private; 13yr is short and strategy is average

Context

Founded by the core creators of Apache Spark at UC Berkeley. Popularized the 'lakehouse' concept by merging data lake and data warehouse. Strengthened the AI stack by acquiring MosaicML for $1.3B. 2024 ARR over $3B, valued at $62B. IPO observers expect post-2026.

TelesTeles' Insight

Databricks's strength is that it found the right answer to 'how do you turn open source into a business?' Apache Spark is free for anyone to use, but enterprise operations are hell. The real product is the managed service that absorbs that operational burden. Acquiring MosaicML for $1.3B shored up the AI stack and split the market with Snowflake. Berkeley researchers scaled to a $62B valuation. Pre-IPO, ARR was already over $3B (by 2026, a $7B+ revenue run rate and a $190B valuation). The moment 'lakehouse' became the industry's standard term, the match was decided.

12 Criteria (50/60)

StrengthsA Market, B Problem, C Solution, D Team, E Traction, F Moat, G Revenue, K Timing, L Insight

AMarket5/5

Data + AI — over $500B combined. Every enterprise is a customer.

BProblem5/5

Analytics and machine learning live on separate stacks. Data movement cost is massive.

CSolution5/5

The lakehouse. One stack runs both. Weaponizes the Spark open-source asset.

DTeam5/5

Berkeley-bred core developers of Spark. Technically, no one can beat them.

ETraction5/5

ARR over $3B. More than half the Fortune 500 are customers. Splits the market with Snowflake.

FMoat5/5

Dominance over the Spark ecosystem + standardization of the lakehouse. The community is the biggest moat.

GRevenue5/5

Consumption-based billing + enterprise contracts. Grows automatically as customers expand.

HFinancial4/5

Loss-making from massive acquisitions + R&D. 50%+ revenue growth keeps it off the critics' list.

IEfficiency3/5

Heavy data-center costs. Sales expenses are high.

JExit4/5

IPO expected. Could become a $100B-class IPO.

KTiming5/5

Perfectly overlapped the AI/ML explosion. An era where data is the fuel.

LInsight5/5

The right answer to commercializing open source is to be the 'platform,' not the 'tool.'

See other cases

View all cases

NGRAPH — AI ADOPTION SUPPORT

We start AI adoption by understanding how your business is built

At NGraph, engineers come into your workplace, look at your work, decisions and the data you use, and then build AI in. We don't stop at teaching you how to use it: we leave AI in a form your team keeps using.