Menlo, Unusual back Oliver AI's pre-seed for agent-ready database
What's the deal? Oliver AIDealroom has a profile for this one. Try Dealroom → has raised a pre-seed round from Menlo Ventures and Unusual VenturesDealroom has a profile for this one. Try Dealroom → to scale OliverDBDealroom has a profile for this one. Try Dealroom →, an analytical data platform built for agentic AI. The company announced the funding on August 26, 2026.
Why now? Enterprise analytics infrastructure wasn't designed for AI agents, which generate continuous, machine-speed workloads. As deployments scale, that infrastructure becomes both a performance bottleneck and a cost problem, putting the promised return on enterprise AI at risk.
What's the endgame? OliverDB lets AI agents investigate enterprise data in real time using a fraction of the compute of existing systems. In tests using ClickBench query shapes, Oliver ran hundreds of times faster than ClickHouse on CPUs and thousands of times faster with GPU execution.
"This is not a 25% or 50% improvement to existing analytics infrastructure. It's a complete step change," said Praneet Sharma, co-founder of Oliver AI. He said a multi-petabyte environment needing thousands of servers today can collapse to a couple of servers, or even a single GPU.
The product: Beyond speed, OliverDB adds governance and observability across enterprise databases and Model Context Protocol (MCP) servers. Enterprises can set per-agent policies, limit data and actions, and keep an attributable record of every interaction. A separate "model swarm" runs many small, specialised models in parallel to lower token spend and reduce reliance on frontier models.
Investor view: Both backers frame Oliver as a new infrastructure category. "Snowflake and Databricks helped define new categories of data infrastructure for the cloud era," said Tim Tully, partner at Menlo Ventures. "Agentic AI represents another fundamental shift."
John Vrionis, founder at Unusual Ventures, said the challenge in enterprise AI is "giving AI systems meaningful access to valuable enterprise data without sacrificing control, reliability, or trust."
The signal: The names on the cap table matter here. Two established firms backing a pre-seed points to growing investor conviction that agentic AI needs its own infrastructure layer — not just faster models, but faster, governed data systems built for how agents actually work.
Image credit: Cory M. Grenier
