Databricks & the Future of the Firm
Summary
- Databricks is driving two major shifts in the workplace: agentic automation and employee empowerment, allowing those closest to problems to act without waiting for hierarchical approval.
- The company has surpassed a $7 billion run rate with 80% year-over-year growth, growing from a Berkeley research project to a platform where agents now launch over 17 million databases daily.
- Databricks evolved through four technology S-curves: Spark (fast in-memory data processing), Lakehouse (merging data lakes and warehouses for unified storage), Lakebase (separating storage from compute for operational databases), and agents.
- The Lakehouse architecture unifies structured analytics and unstructured storage, while the newer Lakebase product integrates serverless Postgres to enable real-time transactions and analytics without fragile pipelines.
- Ali Ghodsi, Databricks CEO, proposes a "System of Record for Agents" (SORA) to replace human-centric ontologies, ensuring AI agents have access to authoritative data, defined permissions, and evaluation metrics.
- Databricks tools like Genie and Unity Gateway allow non-technical users to query data, build agents, and manage AI identities, budgets, and guardrails across the organization.
- The article argues that AI will trigger a "Second Managerial Revolution" by automating information routing, shifting human roles from execution to management, supervision, and strategic judgment.
- The concept of the "Sovereign Firm" is introduced, where companies build sustainable competitive advantages by owning their data loops, policies, and learned exceptions rather than relying solely on rented AI models.
- Jack Dorsey's Block is cited as an early adopter testing a model that reduces hierarchy by using a continuously updated "world model" to coordinate work and improve information flow.
- As agents automate routine tasks, the value of human workers will shift toward solving unsolved problems, defining new processes, and making high-stakes decisions that require intuition and ethical judgment.
- Databricks rejected a $6 billion acquisition offer years ago to pursue the larger opportunity of building the foundational operating system for the future of the firm.
- The long-term vision suggests that in the coming decades, human work will focus on "what hasn't been figured out yet," as agents handle known procedures and humans direct the creation of new value.
How agents and new database technology will change the shape of the firm—and what we do at work all day
- Keywords
- Growth