What matters in AI.

Subscribe

PhoenixAI got $80 million to make a database for AI agents

PhoenixAI tells that AI agents send many unplanned queries at the same time, and large vendors are too slow.

PhoenixAI, the trade name of CelerData, got $80 million this year. It makes a database that gives answers about company data in real time. Ross Armstrong of PhoenixAI tells at Cloudera EVOLVE26 that agents can send queries to some databases at the same time. PhoenixAI is a database partner of Cloudera and an alternative in the Cloudera marketplace.

What is new

PhoenixAI has a columnar engine. It works with data that has a set shape and data that has none. It finds how to answer each query when the query comes in.

The proof

The source gives no speed tests, no customer numbers and no prices. Armstrong tells that the large database vendors cannot give answers in real time correctly. This is the view of Armstrong.

What is planned

AgentBase is a plan of PhoenixAI. It is a context environment that limits each agent to the data that the agent must have. Armstrong tells that it is not necessary to use PhoenixAI to operate agents. PhoenixAI Anywhere is a self-managed edition. It operates in the cloud of the customer, in the buildings of the customer or in air-gapped areas.

Insight: 2 vendors offer the gate between agents and company data

AgentBase is a plan of PhoenixAI to limit the data that an agent sees. DataHub tells that its tool gives each agent only the data groups that it uses. The 2 vendors make a layer between agents and data, and this reads as a race to control the gate. No vendor gave an external test. The first layer that you use can be hard to change.

Sources

  1. Agentic AI architecture demands real-time databasesSiliconANGLE
AI MATTER · NEWS · AI MATTER · NEWS ·1 NOV2026

Posted

Learn the terms in this story

More in Deals

All Deals news