Takeaways from Databricks Data + AI Summit 2026

Research evidence excerpt

Takeaways from Databricks Data + AI Summit 2026

s directly on the lakehouse with tolerance for 12k+ QPS all while

maintaining sub-second response times. We think this becomes strategically

important as agents proliferate across the org chart and run concurrently in a

production setting as these workflows will increasingly depend on low latency

responses at the data layer for agents to keep iterating.

An expanded Genie portfolio enables enterprise agentic workflows. Genie One

was positioned as the front end experience for business users to reason across

enterprise data, while Genie Agents, Genie Code, and Genie Zero Ops extend that

concept into agent creation, developer workflows, and autonomous data

operations respectively. Genie Code more specifically was noted to be particularly

capable across data engineering as well as ML workflows, consistent with

Databricks' unique heritage across these areas and powered by what

management referred to as the secret sauce of Genie Ontology. Unity AI Gateway

fits into this same vision as the control plane to manage security, model choice,

and cost as these workflows move into production. This likely becomes important

because the agent system of record thesis demands both strong underlying

data/context as well as governed usage at scale, although we note that the

agentic control plane remains a widely contested battleground across various

infrastructure software vendors. We believe Databricks' right to win stems from

the context and control it can offer, especially in the largest of enterprises.

Several infrastructure and ecosystem updates bolster platform value. Within

Lakeflow, ZeroBus Ingest is a fully managed ingestion layer that management

described as wire-compatible with Kafka and able to land streaming data directly