PhoenixAI Bets 80 Million Dollars That AI Agents Will Break Every Database Built Before Them

Quick Facts

• PhoenixAI, formerly CelerData, raised an $80 million Series B in June 2026 and rebranded to focus exclusively on real-time databases for AI agents.
• The company handles over 10,000 concurrent queries at sub-second latency across billions of rows, and streaming data becomes queryable within seconds of arrival.
• PhoenixAI announced a foundational database partnership with Cloudera on September 24, 2026, giving enterprise customers one-click deployment inside Cloudera's cloud environments.

AI agents are breaking enterprise databases. That is the argument PhoenixAI made at Cloudera EVOLVE26 in New York City on October 27, and the company has the architecture receipts to back it up.

Ross Armstrong, SVP of Global Sales at PhoenixAI, told theCUBE at the event that existing platforms cannot keep pace. "There was a real gap in the market for real-time analytics, and to be able to not only ask a question but get the response back in real time was something that all the database players that are out there, even the big names in the market, couldn't do well," Armstrong said.

The Problem With Legacy Analytical Databases

Human analysts run dashboards on schedules. AI agents do not. They fire queries unpredictably, in parallel, and they expect answers fast enough to keep a downstream workflow moving. Armstrong framed the challenge plainly: "I think what the biggest demand is that we don't know what the demand is going to be from those agents. You could have thousands and thousands of queries hitting at the exact same time, but you still need that real speedy response."

Rick Underwood, who joined as president of PhoenixAI at launch, put the architectural mismatch in starker terms. "Most of today's analytical databases were architected for a world that no longer exists, where humans ran dashboards on flat tables and complexity was someone else's problem," he said.

Andy Ye, co-founder and chief product officer, made the production stakes concrete. "Concurrency at this level isn't a benchmark figure, it's what a single production agent deployment generates on an ordinary day."

What PhoenixAI Built Instead

The company rebranded from CelerData in May 2026 and ships a columnar engine that handles sub-second SQL with complex joins on live data without precomputation. It ingests mutable streaming data from Kafka and Flink CDC, reads lakehouse tables in Apache Iceberg, Delta Lake, and Hudi in place, and joins real-time and historical data in a single query.

The platform handles over 10,000 concurrent queries at sub-second latency across billions of rows. Streaming data becomes queryable within seconds of arrival. The self-managed edition, PhoenixAI Anywhere, runs in a customer's cloud, on-premises, or in air-gapped environments.

PhoenixAI is also the original creator of StarRocks, the open-source analytical database now hosted by the Linux Foundation under the Apache 2.0 license.

AgentBase and the Governance Question

Speed alone does not close enterprise deals. Governance does. PhoenixAI is building AgentBase, a purpose-built context environment that restricts agent access to only the data each agent needs. Armstrong described the value directly: "You can design that playground for those agents to be protected and to be governed but also to give them access to the data that you need."

The platform includes role-based access control, SOC 2 standards, encryption, and data sovereignty controls. Those features address a problem that has stalled many early agentic deployments, where agents pulled data they had no business touching.

The Cloudera Deal and the Money Behind It

PhoenixAI joined Cloudera as a foundational database partner in September 2026, letting enterprise customers spin up a managed PhoenixAI cluster in a few clicks through the Cloudera Anywhere Cloud Marketplace. The integration eliminates mandatory data denormalization and ETL pipelines by querying lakehouse data in place.

Cloudera described the rationale in direct terms: "Through Cloudera's collaboration with PhoenixAI, we can offer them a real-time database built the way agents need it to work, with the governance our customers expect."

The $80 million Series B, announced June 11, 2026, was led by Sky9 Capital with participation from Atypical Ventures, Olive Technology Ventures, and prior investors. PhoenixAI also won AI-Powered Enterprise Agent Solution of the Year at the 2026 Tech Ascension Awards.

The company is placing a direct bet that as agentic workloads move from prototypes to production at scale, the database sitting underneath them will become a competitive variable, not a commodity. Armstrong's warning to enterprises building on legacy systems was unambiguous: if agent demand cannot be forecast, the database has to be ready for anything.