🤚 The Open-Palm Valuation Event

Databricks, the company that started life in 2013 as enterprise data analytics software and has since rebranded itself as the backbone of AI infrastructure, just closed a funding round that values it at $188 billion. The approximately $3 billion raise was led by Coatue, a firm that apparently looked at the current state of the AI industry — burning cash at historically unprecedented rates, generating revenue at historically ambiguous ones — and said, “yes, but what about the company that sells the storage?”

For context, here is the company’s recent valuation trajectory:

• December 2024: $62 billion
• September 2025: $100 billion
• February 2026: $134 billion
• July 2026: $188 billion

That is a tripling in eighteen months. If your home appreciated at the same rate, your mortgage broker would have called the SEC.

👐 The Two-Handed Infrastructure Paradox

What makes Databricks fascinating is not that it builds AI models. It does not. What it builds is everything around the AI models — the data lakehouse, the analytics layer, the governance platform — and then it tells its 3,000 software engineers to use open-weight models instead of paying Anthropic or OpenAI for the privilege.

CEO Ali Ghodsi recently published internal benchmarking research showing that open models — and GLM 5.2 in particular — can now “handle even the highest level of task difficulty” at a fraction of the cost of proprietary alternatives. The company has introduced a suite of products that sound like they were named by a fantasy novelist with an MBA:

• Lakebase — an AI agent database
• Unity — an AI gateway
• Omnigent — an agent management tool

The strategy is brilliantly simple: let OpenAI and Anthropic spend $100 billion a year training frontier models, wait six months for an open-source equivalent to appear, and then sell the infrastructure to run it. It’s the business model of a man who sells umbrellas outside a cathedral where the priests are praying for rain.

🌿 The Gentle Awakening

There is something quietly profound about a $188 billion company whose entire thesis is that the most expensive AI models are not worth what they cost. Databricks is not betting against AI. It is betting against the margin structure of AI — which, if you have been following the industry, is the one number that nobody wants to discuss at earnings calls.

The big labs spend billions training models. They charge pennies per token. They lose money on every API call and plan to make it up in volume. Databricks looked at this arrangement and concluded that the winning move is to be the neutral party — the Switzerland of the AI stack, except Switzerland charges for parking.

Meanwhile, every traditional enterprise software company is watching Databricks triple its valuation while they struggle to explain why their own AI features are not generating the revenue they promised on last quarter’s investor call. The answer, of course, is that enterprises do not want AI features bolted onto their CRM. They want infrastructure — and Databricks has been selling infrastructure since before “AI” replaced “cloud” as the mandatory keyword in every pitch deck.

👑 The Gold-Leaf Market Thesis

At $188 billion, Databricks is now worth more than Goldman Sachs. It is worth more than the GDP of Hungary. It is approaching the valuation of companies that have things like “profit” and “dividends” and “products you can hold in your hand.”

But the market is not valuing Databricks on what it earns today. It is valuing Databricks on the assumption that every company on Earth will need an AI data platform, that open-weight models will continue to close the gap with proprietary ones, and that the real money in the gold rush was always in selling the picks and the shovels and the cloud-native lakehouse with built-in governance compliance.

The company has not disclosed its revenue. This is either a sign of supreme confidence — the numbers speak for themselves, we just won’t tell you what they are — or it is the kind of strategic opacity that makes investors feel like they are in an exclusive club rather than a speculative bet.

Either way, Databricks has achieved something rare in the current AI landscape: it has become indispensable without ever training a frontier model, publishing a benchmark, or getting subpoenaed by the Senate. In an industry defined by main-character energy, Databricks is the supporting actor who somehow owns the theater.

“Our valuation tripled in eighteen months and we didn’t even have to build a foundation model. We just sold the foundation.” — The Slap of Wisdom Infrastructure Desk, quietly adding ‘Lakehouse’ to its LinkedIn headline