Signals Inbox·July 23, 2026·AI Infrastructure

Is Databricks really worth $188B today?

Databricks can plausibly grow into a $188 billion valuation, but the price already assumes another year of exceptional growth and margins the company has not yet disclosed.

Q1What happened to Databricks’ latest valuation?

Databricks’ latest valuation currently rests on a signed term sheet, so the headline price is serious but still provisional.

Databricks recently announced that existing investor Coatue would lead the strategic financing, with additional new and existing investors expected to join. The Wall Street Journal reported that the investment would total about $3 billion. The proposed price stands 40% above the $134 billion valuation agreed in December 2025.

The speed is unusual even by late-stage technology standards. Databricks was founded in 2013, reached $38 billion in 2021 and then moved from $62 billion to the current proposed price in roughly 19 months. The latest step also leaves the company valued at about twice Snowflake’s current public-market value.

The deal can still change before closing. We should treat the figure as a real market price with more uncertainty than a completed financing or a liquid stock-market valuation.

Databricks valuation history

Q2Has Databricks’ valuation outrun the business?

Yes, Databricks’ valuation has risen faster than its revenue, so investors are paying a larger multiple than they did in late 2024.

At the $62 billion Series J price, Databricks expected to cross a $3 billion revenue run-rate. That implied about 20.7 times revenue.

From those two points, the valuation rose 203% while the revenue base rose about 130%. The revenue growth explains most of the jump, but investors also expanded the multiple by roughly 32%.

Databricks earned a higher price through exceptional growth. The remaining increase comes from a stronger belief that it can control a large part of enterprise data, analytics and AI spending for years.

Q3Which Databricks revenue number should we trust?

We should currently anchor the valuation to $6.9 billion of annualized revenue, while treating the management-reported figure cautiously.

Databricks disclosed the number to analysts during its latest Data + AI Summit, and CEO Ali Ghodsi discussed it in a CNBC interview. The company said annualized revenue had grown by more than 80% year over year, with AI products contributing around $1.7 billion.

The last detailed figure in a formal Databricks press release was $5.4 billion in February, with growth above 65%. Databricks also reported positive free cash flow over the preceding 12 months and net revenue retention above 140%. That release is a cleaner official baseline, though an older one.

Annualized revenue takes a recent period of customer consumption and projects it across a full year. Public companies usually give investors recognized quarterly revenue, contracted obligations and audited accounts. Databricks has yet to provide a full reconciliation between those measures.

Using the summit figure produces a 27.2-times multiple. Using the last formal company update produces 34.8 times. The newer number better reflects current demand, while the gap between the two calculations shows how much uncertainty remains.

Q4Does any comparable company trade this high today?

Databricks’ revenue multiple is expensive, yet similar companies show that investors sometimes pay this much for exceptional AI growth.

Snowflake currently trades near 17 times annualized revenue after growing 33%. MongoDB trades near nine times after growing 25%. Palantir sits above 50 times after an 85% growth quarter, a 46% GAAP operating margin and a 57% adjusted free cash flow margin.

Private benchmarks tell a similar story. Glean’s $7.2 billion funding price now equals about 24 times its subsequently reported $300 million ARR. AlphaSense recently raised at $7.5 billion after passing $600 million ARR, or roughly 12.5 times. Glean is growing much faster than AlphaSense, which helps explain the gap.

Databricks should trade above the slower public and private companies thanks to much faster growth from a far larger revenue base. Palantir shows that investors will pay more for comparable growth when the margins are exceptional and fully disclosed.

Revenue multiples across Databricks and selected peers

Q5Is Databricks really worth twice as much as Snowflake?

Databricks deserves a premium to Snowflake today, but the current price assumes that the growth gap will remain wide.

Snowflake’s latest quarter produced $1.39 billion of revenue, up 33%, with net revenue retention of 126%. It had 779 customers generating more than $1 million in trailing product revenue. Databricks reported more than 800 customers consuming above that threshold and net revenue retention above 140%.

That comparison is unusually clean. Both companies have reached a similar number of very large accounts, while Databricks is getting much faster expansion from existing customers. Its wider platform also reaches data engineering, machine learning, governance and AI development, areas where Snowflake is expanding as well.

The premium makes sense. The difficult part is its size: Databricks is valued at roughly double Snowflake even though its revenue base is only around one-quarter larger. Investors are pricing several more years of superior growth before those years have happened.

Q6Is Databricks still accelerating at this scale?

Databricks is currently accelerating at a scale where most software companies are already slowing down.

The company crossed a $4 billion run-rate in September 2025 with growth above 50%, then moved past $4.8 billion in December with growth above 55%. It reached $5.4 billion in February with growth above 65%. Its latest summit update pushed the reported growth rate above 80%.

The absolute additions also became larger. Databricks added about $800 million of annualized revenue between the first two updates, another $600 million by February and roughly $1.5 billion by the next summit. Across about nine months, the run-rate increased by 72.5%.

Both the growth rate and the dollars added rose across four consecutive disclosures. An IPO filing may eventually show more quarter-to-quarter variation than these run-rate snapshots, but the pattern available now is exceptionally strong.

Q7Is AI taking over Databricks’ revenue?

AI is already a large Databricks business and currently contributes about one-quarter of total revenue.

Databricks reported $1 billion of AI revenue against a $4 billion total run-rate in September 2025. By February, the figures had moved to $1.4 billion against $5.4 billion, followed by $1.7 billion against the latest company total. Those combinations imply AI shares of 25%, 25.9% and 24.6%.

The share barely moved. AI revenue grew about 70% between the first and latest disclosures, while total annualized revenue increased by roughly 72.5%.

The boom is spreading through the whole platform. AI agents create more queries, which drives consumption in data preparation, storage, governance and analytics alongside the dedicated AI products. The valuation rests on a broad data platform that earns more as enterprise AI use expands.

Q8Are Databricks customers unusually sticky?

Databricks’ customer economics are strong enough to support a premium valuation today.

Net revenue retention above 140% means the average existing customer group increases annual spending by more than 40% after churn and contraction. The number of customers consuming above $1 million annually rose from more than 650 in September to more than 700 in December and more than 800 in February. Over 70 customers were already above a $10 million annual run-rate.

The company now says more than 20,000 organizations use its platform, including 70% of the Fortune 500. These are company-reported figures, and the large-account progression lines up with the revenue acceleration.

The spending pattern is easy to understand. A customer may begin with data engineering, then add warehousing, machine learning, governance and AI applications. Each extra workload raises consumption and makes a later migration more disruptive.

Customer retention is among the strongest evidence in the whole valuation case. High retention at a multibillion-dollar revenue base is hard to explain with temporary AI experiments alone.

Q9What keeps customers from moving off Databricks?

Databricks’ moat comes from years of accumulated workflows, permissions and data dependencies, which are painful to rebuild elsewhere.

The company was founded by the original creators of Apache Spark and later built or helped create Delta Lake, MLflow and Unity Catalog. Those open-source projects gave Databricks early distribution among data engineers and machine-learning teams.

Open source also makes the platform easier to adopt without forcing customers into a proprietary format. Databricks open-sourced Unity Catalog and supports both Delta Lake and Apache Iceberg, allowing other engines to work with governed data.

Customers become harder to move later. Once a company has built pipelines, access rules, models, monitoring and applications around the same platform, leaving requires a long technical migration across several teams. Replacing Databricks involves far more work than swapping one database.

Competitors can reproduce individual features. Rebuilding a customer’s full operating setup is much harder, and all that work keeps customers around.

Q10Can Amazon, Microsoft or Google squeeze Databricks?

The cloud giants can pressure Databricks’ pricing, but none currently offers a simple replacement for the whole platform.

Amazon, Microsoft and Google own the underlying infrastructure, control large enterprise sales channels and can bundle databases, analytics and AI services. Every extra model call or compute job also leaves Databricks exposed to infrastructure costs set by its partners.

Databricks has reduced that risk through a multi-cloud product, open formats and deep partnerships with all three providers. Customers can keep a more consistent data layer while using different clouds, models and tools. Unity AI Gateway extends that idea by routing and governing access to models from several providers.

The relationship remains uncomfortable. The hyperscalers benefit when Databricks drives more cloud consumption, while their own products compete for the same software margin. Pricing pressure should remain steady, especially in warehousing and operational databases.

Cloud providers can copy features faster than they can untangle the pipelines, permissions and applications already running inside a customer’s Databricks environment. That gives the company room without guaranteeing permanent pricing power.

Q11Is the market expanding fast enough for Databricks?

The market is expanding fast enough to support a much larger Databricks, although the broadest AI spending totals overstate its direct opportunity.

Gartner currently forecasts worldwide AI spending of about $2.59 trillion for 2026, up 47%. Much of that total goes to chips, servers, cloud infrastructure and services outside Databricks’ product line. A narrower Gartner estimate puts spending on AI models and platforms at $64 billion, up 63.4%, with AI platform spending alone rising 36.9%.

Enterprise buying surveys support the demand story. A recent RBC survey covered more than 100 CIOs and technology leaders. Every respondent had an AI budget, and 91% were creating new budgets. More than half already had AI in production, while another 35% expected production use within six months.

Those budgets flow toward Databricks when companies need cleaner data, governance, model access and cost controls. The need is especially clear as agents create more queries and less predictable bills.

The category is large and moving quickly. Databricks still has to win a meaningful share of those budgets, but there is plenty of room for the company to grow.

Q12Are Databricks’ margins good enough?

Margins are currently the weakest part of the valuation case because Databricks says it generates cash without showing how much it keeps.

The company repeated the positive free cash flow claim across three consecutive revenue updates. That consistency makes the claim credible, but investors still lack the cash amount, gross margin, operating margin and stock-based compensation.

Ali Ghodsi has also said that margins are shrinking as AI agents generate more queries and Databricks pays for additional model usage and infrastructure. Revenue can rise quickly under that consumption model while each new dollar carries a lower gross margin.

Snowflake and Palantir publish the numbers Databricks has yet to provide. Snowflake guides to a 75% non-GAAP product gross margin and a 23% adjusted free cash flow margin. Palantir recently reported a 46% GAAP operating margin and a 57% adjusted free cash flow margin.

Until an IPO filing shows Databricks’ gross margin and how much cash the business keeps, investors are paying for growth quality they cannot fully inspect. That disclosure gap keeps us from calling the price fully justified today.

Q13Is Databricks spreading itself too thin?

Databricks is expanding into too many markets to call every move low-risk, yet the product logic still hangs together.

The company acquired MosaicML to move deeper into generative AI, bought Neon to create Lakebase, launched the Lakewatch security product and later agreed to acquire Panther. It has also introduced Genie One, real-time analytics, CustomerLake, LTAP and OpenSharing.

During one recent summit cycle, Databricks announced an AI coworker, a real-time system, a customer-data platform, a new transactional architecture and a security acquisition. That is a lot. Each product can still run on the same governed data, so customers can add it without rebuilding their foundation.

Execution becomes harder as the buyer set expands. Data engineers, database developers, security teams, marketers and business users have different priorities and sales processes. Product breadth can also create a confusing platform if integrations lag behind announcements.

The next useful proof will come from customer adoption across several modules. Another crowded launch calendar would tell us much less.

Q14How much revenue would justify the valuation?

Databricks can grow into the proposed valuation fairly quickly if growth stays above 35%, but a normal software slowdown would leave the price exposed.

A 25-times multiple requires $7.52 billion of revenue, only 9% above the latest baseline. A 20-times multiple requires $9.4 billion. Reaching a more ordinary 15-times multiple needs $12.53 billion.

Revenue needed to support a $188B valuation

Q15What has to go right, and what would break the valuation?

The price works if Databricks stays above roughly 40% growth for two years; it breaks quickly if growth settles near 20% to 30% before margins become visible.

Two years of 40% growth would take annualized revenue to about $13.5 billion. At 15 times revenue, that business would be worth roughly $203 billion. Two years at 50% would produce around $15.5 billion and support about $233 billion at the same multiple.

That path also requires net retention to remain above roughly 130%, Lakebase or Genie to become major products, and AI-related infrastructure costs to stay under control. The case can work with only a few new products becoming billion-dollar businesses, provided the revenue is durable and reasonably high-margin.

The bear case needs only one bad turn. One year of 30% growth would produce about $9 billion of annualized revenue. Applying Snowflake’s current multiple would yield a value near $152 billion. At 20% growth and a 15-times multiple, the result falls near $124 billion.

Databricks can execute well and still disappoint investors. A move toward normal software growth before the revenue base catches up would erase a large part of the current price.

Q16So, is Databricks really worth that price today?

At $188 billion, Databricks is aggressively priced but still plausible, with part of the valuation resting on growth that has yet to arrive.

The company has built a rare combination: multibillion-dollar scale, accelerating growth, deep enterprise adoption and strong customer expansion. Its open-source roots and the amount of work customers build around the platform give it more staying power than a typical AI application startup.

The price is roughly 27 times the latest annualized revenue, compared with about 17 times for Snowflake. Palantir shows that the market can pay far more for similar growth, backed there by exceptional disclosed margins that Databricks has yet to match.

Our call is that the valuation is aggressive but plausible. It becomes reasonable if revenue passes roughly $9 billion soon, growth stays above 40% and the eventual financial statements show healthy gross margins. A slowdown toward 30% before those conditions appear would put the fair value below the current round.

Databricks may grow into the price quickly. Right now, investors are paying for roughly one more year of outstanding execution before they have the audited evidence.