Databricks Raises $1B at $100B Valuation to Build Lakebase

Databricks has quietly closed a new, wildly oversubscribed funding round worth about $1 billion at a $100 billion valuation, according to sources. The deal was co-led by Thrive and longtime backer Insight Partners, the same firms that led the company's prior round. This primary raise follows two 2025 secondary windows that let employees sell portions of their holdings.

Unlike some rounds that shore up operating cash, Databricks says it doesn’t need funds for day-to-day operations. CEO Ali Ghodsi told TechCrunch the fresh capital has a specific purpose: accelerate Lakebase, a Postgres-based database tuned for AI agents, and scale Agent Bricks, Databricks’ agent orchestration platform launched earlier this year.

Why Lakebase and Agent Bricks matter

Databricks argues the database market is a $105 billion opportunity stuck in legacy dynamics, with incumbents like Oracle dominating for decades. Ghodsi points to a striking shift: databases created by AI agents rose from 30% last year to 80% this year, and he forecasts nearly all new databases will be agent-created within a year. That changes who the "user" is—and how systems must be designed.

Lakebase differentiates by separating compute and storage so organizations can let fast, short-lived agent workloads spin up many databases without incurring crippling compute costs. In plain terms: treat storage like cheap shelf space and compute like on-demand machines that scale independently. That design is meant to let agents create and discard environments cheaply and at speed.

What this means for enterprises

If you run data platforms, DevOps, or enterprise applications, Databricks’ move is a signal: agent-first workloads will require new cost models, governance, and orchestration patterns. Expect pressure to rethink licensing, observability, provisioning speed, and how to prevent runaway agent spend or data sprawl.

• Assess where agent-driven databases unlock business value, from HR automation to customer personalization.
• Model separated compute/storage economics to control costs when agents spin up many short-lived databases.
• Design governance, access controls, and monitoring for machine users so agent actions are auditable and safe.

Databricks isn’t just chasing research-level AGI; it’s betting on practical, repetitive agent tasks—onboarding, benefits Q&A, and workflow automation. Those are the kinds of workloads that scale across enterprises and affect productivity and costs in measurable ways.

How to respond now

Short-term pilots, paired with cost and governance guardrails, are the fastest way to learn. Think small: a single HR or support workflow can reveal agent cost behavior, data lifecycle needs, and integration gaps. Over time, expand to standardized agent database patterns and centralized observability.

At QuarkyByte we analyze vendor architecture trade-offs, model financial implications of separated compute/storage, and build pragmatic adoption roadmaps so organizations can pilot agent workloads safely and economically. The Databricks move accelerates a shift that teams should evaluate now—not later.