The Quiet Infrastructure Story: Databricks And Microsoft Extend Into The 2030s

Most of the attention in AI goes to the companies building the models. But the harder, less glamorous question is what happens after the demo — when a bank, a carmaker, or a retailer tries to run AI on its own messy data, every day, at a price that makes sense. That's where this week's news lands.

On July 23, Databricks said it would extend its long-running partnership with Microsoft into the 2030s. The San Francisco company, whose platform helps organizations pull together scattered data and build AI on top of it, will run more of its own operations on Microsoft's Azure cloud — and, notably, lean harder on the custom chips Microsoft has been quietly designing in-house.

What actually changed

The headline is the chips. Databricks will expand its use of Azure Cobalt, Microsoft's own Arm-based processors, for the most demanding jobs: data-heavy analytics and the new wave of "agentic" AI workloads, where software doesn't just answer a question but carries out a task. Databricks already runs on the current Cobalt 100 and plans to move to the next generation, Cobalt 200, which Microsoft says delivers meaningfully better performance and turns on memory encryption by default.

That last detail matters more than it sounds. Faster, purpose-built silicon tends to mean more work done per dollar and per watt of electricity — and encryption on by default means security that doesn't depend on someone remembering to switch it on. For the enterprises actually footing the bill, those are the differences between an AI project that pencils out and one that quietly gets shelved.

The deal runs both directions. Microsoft will keep weaving Databricks' tools deeper into its own products — including Genie, Databricks' conversational analytics assistant that lets people ask questions of their data in plain language, and the Unity AI Gateway, which helps companies govern their models and keep costs in check. "With Databricks deepening its investment in Azure Databricks and Azure Cobalt-powered infrastructure, customers will benefit from greater performance, efficiency, and scale for their most demanding workloads," said Judson Althoff, CEO of Microsoft's Commercial Business.

Why a plumbing story is worth telling

It's tempting to file this under corporate housekeeping. But step back and it's really a story about AI becoming ordinary — in the best sense. The exciting part of any technology is the breakthrough; the part that changes lives is when the breakthrough turns into reliable infrastructure that everyone can use without thinking about it. Electricity mattered most once it was just there in the wall.

Custom chips, governed data, cost controls, encryption by default — this is the unglamorous machinery that moves AI from impressive demos into the everyday tools that help people at work do more with less friction. Two of the industry's most capable builders pointing their engineering at that machinery, over a ten-year horizon, is a good sign for the companies and workers who will ultimately depend on it.

The scale behind the deal

Databricks isn't a small player making a hopeful bet. Its platform is used by more than 20,000 organizations worldwide, including roughly 70% of the Fortune 500. And the partnership news arrives just after the company signed off on a new strategic funding round — reported to total around $3 billion and led by existing investor Coatue — that values Databricks at $188 billion, one of the largest valuations any private company has ever carried. That round is expected to close later this summer.

Put those two facts together and the direction is clear: one of the most widely used data platforms in the world is committing its next decade to Microsoft's cloud and Microsoft's chips, at the exact moment enterprise AI is moving from experiment to everyday operations. The next chapter of AI is being written as much in silicon and data pipelines as in the models that grab the headlines — and this is what steady, real-world adoption actually looks like.

This article is based on Databricks' and Microsoft's July 23, 2026 announcements and contemporaneous reporting. Details of the funding round are developing and figures are as reported at the time of writing; valuations and terms may change before the round closes. This is informational only and is not financial, legal, or investment advice.

Sources: Reuters; Databricks and Microsoft newsrooms; Bloomberg; TechCrunch; SiliconANGLE.