Prat
Moghe.

The fourth data company.

In October 2024, Promethium's founder Kaycee Lai gave up the CEO title and moved to Chief Strategy Officer. The person who took the chair, Prat Moghe, had already built two data companies that got acquired and helped run a third. Founders do not usually hand the keys to a stranger. They usually hand them to the person who has done this before.

Promethium, which had raised roughly $34.5M across its rounds, positions itself in the crowded corner of enterprise software called the data fabric. The pitch is short: enterprises never consolidate their data, so the fabric has to come to the data instead of the other way around. Federated queries, active metadata, natural language on top. Moghe's public quote on joining was almost boring in its precision: data fabric is recognized as innovative, and Promethium makes it a practical reality by deploying in minutes rather than the months a bespoke build usually takes.

The company employs 34 people. Moghe once ran a 400-person team at Netezza. The ratio is not an accident. Small teams with veteran operators are how large software categories quietly get built.

To get to Promethium, Moghe took a route that reads like a decade of the enterprise software industry compressed into a resume. Bell Labs, where a lot of people who now run infrastructure companies did their first serious work. A PhD from UCLA in Electrical Engineering. Then Tizor, a startup he founded that did behavioral auditing for data compliance. Tizor was acquired by Netezza. The interesting bit is what happened next: he stayed, and rose to Senior Vice President of Strategy, Products and Marketing at Netezza, running that 400-person team, in time to help sell the whole company to IBM in 2010 for $1.7 billion. There is a version of this story where he cashes out and buys a boat. He founded another company instead.

That next company was Cazena, a cloud data-as-a-service platform. Cazena eventually got acquired by Cloudera, at which point Moghe became EVP and General Manager for Cloud at Cloudera. Somewhere in there he also served as president of TiE Boston from 2012 to 2014 and joined the TiE Board of Trustees, which is how the Boston enterprise-software community tends to sort itself: the people who have exited become the people who mentor.

Then Promethium called.

The interesting thing about a CEO change at a small startup is that everything about the company that mattered on Monday still matters on Tuesday. The engineering roadmap. The customer commitments. The product truth. What actually changes is the way capital, talent, and time get spent. Moghe's opening letter to customers was titled, in earnest, "Democratizing AI and Data One Question at a Time." Which sounds like a slogan until you read the letter, at which point it turns out to be a product description. Ask your data a question. Get an answer. Do not spend six months integrating it first.

Promethium sells into an environment that has arguably never been more chaotic. The average enterprise now runs its data across Snowflake, Databricks, on-prem warehouses, five different SaaS applications and a lake in AWS. Half the AI industry is trying to convince these enterprises to move all of that data into a new store. The other half, Promethium included, is convinced they won't. Federated query engines, natural-language interfaces, agents that plan across sources: the bet is that the interface layer will win because the underlying storage layer never converges.

Moghe has been placing versions of this bet since Tizor. In an interview with insideAI News back in 2021, when he was still running Cazena, he was already talking about the collision between the desire for self-service data and the reality of governance in a regulated enterprise. The tools have changed. The problem has not.

What has changed is that the natural-language interface finally works well enough to be pointed at production data. Promethium's product listing sits squarely in the middle of the current text-to-SQL wave, and Moghe is one of the few CEOs in that wave who was building augmented data catalogs before the phrase "GenAI" existed. He speaks at Databricks' Data + AI Summit. He writes for DATAVERSITY and Entrepreneur. His byline is one of the older ones on the topic, in the way that a Wall Street analyst's byline can be old on any given trade.

His public writing is notably free of the language most CEOs use. Not "category-defining." Not "paradigm shift." Just: here is the customer problem, here is the mechanism, here is why we can do it faster than the incumbent. His letter announcing his arrival at Promethium reads more like an internal memo than a press release, which is probably why the company published it.

The other quiet detail worth noting is the shape of his career: four companies, two acquisitions, one PhD, and a lot of years spent as the number-two or number-three name on the org chart of companies that later became famous. He is not the founder-CEO who gets on magazine covers. He is the operator who takes the company through the middle chapter, when it needs to become a real business.

Promethium is his fourth data company. If the pattern holds, it will not be his last one that gets acquired.