Ocient Raises $100 Million to Run AI Agents Inside the Database, Betting Governance Beats Model Access

Ocient has closed a $100 million Series C co-led by BOKA Capital and OCA Ventures, bringing total funding to $260 million. The Chicago company did not disclose a valuation.

Every existing investor took part: Greycroft, In-Q-Tel, the venture arm tied to the US intelligence community, Blue Bear Capital, Buoyant Ventures, PSP Growth, Pritzker Group Venture Capital and investor Eron Wolf. Dallas Venture Capital joined as a new backer, and BOKA chief executive John James takes a board seat.

What Ocient sells is a single engine that ingests, stores and analyzes petabyte-scale data without sampling, now with a layer that lets AI agents query the same governed data. The OcientAIQ Unified Data Platform bundles ingestion, access control, analytics, business context and agent skills; OcientAIQ Solutions are industry-specific packs built on top. The architectural argument is that analysis runs inside the engine rather than through a language model, so every result is reproducible and auditable, data movement stays minimal and token costs stay low. The platform also uses forward error correction rather than replication to protect data with less storage overhead.

Governance is the differentiator the company is selling. Access rules, policies and lineage - who may reach which data, which rules apply, where a value came from and how it changed - are enforced on agent activity inside the engine. For a telecom carrier that has to hand call records to investigators, or a security team tracing an indicator across years of logs, being able to prove what an agent saw and under what rule matters more than the model's reasoning.

The scale numbers are Ocient's own and have not been independently verified. The company says one telecom customer ingests more than a million call detail records a second while querying trillions of records at the same time. It describes Snowflake and Databricks as complements more often than replacements, and has not published benchmarks against either.

Growth is where the story gets interesting. Ocient expects revenue to more than triple in the fiscal year ending in January, after doubling the year before. Absolute revenue was not disclosed, so the percentages cannot be sized against a base, and a handful of large telecom contracts can swing a company at this stage. Chief executive John Morris said the growth comes from new national security and telecom customers consolidating cybersecurity data on the platform, and that most of the current pipeline is on-premises as buyers repatriate workloads for cost, sovereignty or data protection reasons. Ocient formalized a national security subsidiary and an India operation earlier this year.

The company was founded in 2016 by Chris Gladwin, who previously built Cleversafe and sold it to IBM for $1.3 billion in 2015, together with Joe Jablonski and George Kondiles. Morris became chief executive last June and Gladwin stayed on as executive chairman. Ocient raised a $42.1 million Series B extension in April 2025.

The bet underneath the round is that agentic AI moves the bottleneck from model quality to provable access. Agents ask far more questions than people, and if every read is mediated by a language model, costs rise, answers stop being reproducible and audit trails dissolve. Ocient is arguing that a narrow, unglamorous tier - extreme ingest rates, tens of petabytes, defense and telecom - is where governance becomes a purchasing requirement rather than a feature. Gartner expects worldwide IT spending to reach $7.35 trillion in 2027, up 15 percent, and a growing share of that is the plumbing that makes AI usable rather than the model itself.