Databricks Expands Microsoft Azure Partnership and Adopts Custom Chips
Miles Bennett
Databricks is extending its Microsoft partnership into the 2030s, scaling up Azure usage and adopting Microsoft's custom Cobalt processors — a landmark endorsement of Microsoft's in-house chip strategy in enterprise AI infrastructure.
What exactly did this deal lock in?
Databricks extended its Microsoft agreement into the 2030s, covering two pillars: scaled-up Azure platform usage + adoption of Microsoft's custom processor Azure Cobalt.
Azure Cobalt — Microsoft's Arm-based chip designed for heavy cloud workloads — will power Databricks' data-intensive and agentic AI tasks.
This means → Databricks is not just renting Azure cloud space. It is running its core operations and data analytics on Microsoft's infrastructure — a deep, long-term lock-in.
Why is Databricks such a valuable customer?
Databricks is one of the most valuable private tech companies in the world. Last week it closed a funding round at a $188 billion valuation, expected to settle later this summer.
Its platform serves over 20,000 organizations globally, including roughly 70% of Fortune 500 companies.
In plain terms = this is not a small client picking Azure. It is the leading enterprise data platform betting its next decade on Microsoft — the kind of deal Azure can headline in an earnings call.
What does each side get?
Microsoft's gain: Databricks' deep adoption of Azure Cobalt gives Microsoft's custom-chip strategy an enterprise-grade reference case. "Even Databricks uses it" is itself a selling point for other customers.
Databricks' gain: Microsoft commits to integrating Databricks' AI capabilities into its own product line, including Databricks' conversational analytics tool Genie — giving Databricks a channel to reach more enterprise buyers.
Microsoft commercial CEO Judson Althoff said: "Customers will get higher performance, better efficiency, and greater scale in the most demanding workloads."
What industry signal does this deal carry?
This reflects a shift: enterprise AI is moving from pilot projects into large-scale infrastructure procurement — top-tier customers are locking in their platform choices for the next decade.
This means → the competition between Microsoft's in-house chips (Azure Cobalt) and Nvidia GPUs is shifting from "whose chip is faster" to "who can lock in more critical customers."
The key thing to watch next: whether Microsoft can use the Databricks deal as a template to push its custom-chip strategy into more enterprise AI workloads and widen its cloud market share.
Content is for reference only, not financial advice.