BREAKING  Promethium ships agentic platform for self-service data at AI scale Meet Mantra - the data-answer agent that returns SQL, lineage & visuals Federated engine reaches 200+ sources - no ETL, no data movement Series A led by Insight Partners - $34.5M raised to date Founded 2018 in Menlo Park by Kaycee Lai Gartner Cool Vendor · SOC 2 Type II certified BREAKING  Promethium ships agentic platform for self-service data at AI scale Meet Mantra - the data-answer agent that returns SQL, lineage & visuals Federated engine reaches 200+ sources - no ETL, no data movement Series A led by Insight Partners - $34.5M raised to date Founded 2018 in Menlo Park by Kaycee Lai Gartner Cool Vendor · SOC 2 Type II certified
Promethium logo
Enterprise Data · Agentic AI · Menlo Park, CA

Promethium.

Ask a question. Get a trusted answer. The Instant Data Fabric that brings the questions to the data - instead of moving the data to the questions.

Founded 2018 ~34 employees $34.5M raised Series A

Above: the Promethium mark - a nod to element 61, promethium, and the company's LinkedIn handle "pm61data." Photographed as filed on the company's press materials.

The Dispatch

A data company betting against moving your data

For fifteen years, the enterprise playbook has been the same: copy everything into one warehouse, transform it, and only then ask it questions. Promethium, a small company in Menlo Park, California, thinks that ritual is the problem - not the solution.

Founded in 2018 by Kaycee Lai and now led by CEO Prat Moghe, Promethium builds what it calls an Instant Data Fabric: a platform that reaches across an organization's scattered systems - cloud warehouses, on-premises databases, SaaS applications - and answers questions where the data already lives. No extract, transform, load. No duplication. The company's wager is that in the age of AI agents, the winning data layer is the one that delivers a trusted answer fastest, with the context to back it up.

That premise sounds simple. Delivering it is not. Behind every plain-English question sits a federated query engine spanning 200-plus sources, a context layer that fuses technical metadata with business logic, and governance strict enough for a regulated enterprise. The user types a sentence. Promethium does the rest.

200+
Data sources reached
$34.5M
Total funding
2018
Year founded
Zero
ETL required
What It Does

Self-service data, wired for AI

Promethium's mission, in its own words, is to "help every business be data driven by enabling every employee to make data-driven decisions in near real time." In practice, that means giving three audiences the same superpower: analysts who need answers without filing a ticket, business users who don't write SQL, and - increasingly - AI agents that are only as trustworthy as the data they can reach.

The company's federated engine queries information across hybrid, multi-cloud environments without centralizing it first. Its 360 Context Engine brings technical metadata together with business logic, so that an answer reflects how the organization actually defines a "customer" or a "region" - not just what a raw column says. And its conversational agent, Mantra, turns a natural-language question into a complete Data Answer: generated SQL you can inspect, the metadata and lineage behind it, a visualization, and governance baked in.

Products & Services
Platform · since 2019

Instant Data Fabric

A federated platform that accesses data across 200+ cloud, on-premises and SaaS sources without ETL, movement, or duplication - queried where it lives.

Agent · 2025

Mantra

A conversational data-answer agent that converts plain-English questions into governed answers - complete with SQL, metadata, lineage and visualizations - shareable into existing tools and AI systems.

Context Layer · 2025

360 Context Engine

Fuses technical metadata with business logic so answers align with how the organization actually works, sharpening relevance across hundreds of sources.

Architecture · 2025

Agentic Analytics Platform

An open agentic architecture purpose-built to deliver self-service data at AI scale - connecting analysts and AI agents to trusted enterprise data.

Most weren't built to keep up. But we're giving data teams a new superpower - the ability to deliver trusted, contextual answers on demand.
- Prat Moghe, Chief Executive Officer, Promethium
The Problem & The Difference

Why the old approach strains under AI

The problem it solves

  • Data lives everywhere, yet answers are still hard to get
  • Every dashboard request spawns three more for the data team
  • Centralizing data first is slow, costly, and duplicative
  • AI agents hallucinate when they lack governed context
  • Business logic gets lost between raw columns and real meaning

How Promethium differs

  • Queries data in place - no mandatory ETL or movement
  • Generates SQL across many sources, not a single warehouse
  • Every answer ships with lineage, metadata and governance
  • Context Engine grounds AI in how the business defines things
  • Built for agents from the start, not retrofitted onto BI

Where much of the market sells the destination - a warehouse or lakehouse to pour everything into - Promethium sells the connective tissue. That places it among data-fabric and virtualization players such as Denodo, Starburst and Dremio, and in quiet opposition to the "move it all first" orthodoxy of the warehouse-centric world. Its differentiator is not just federation but trust: answers you can inspect, trace, and govern, which in a serious enterprise is the price of admission for letting AI touch the data at all.

The Record

Seven years, one thesis

2018

Promethium is founded

Kaycee Lai starts the company in Menlo Park to close the gap between enterprise data and actionable answers.

2019

Early funding, first fabric

Backing from .406 Ventures and Zetta Venture Partners funds the federated data-fabric approach.

2022

$26M Series A

Insight Partners leads a $26M round, bringing total funding to roughly $34.5M to scale customer growth.

2024

Prat Moghe named CEO

The 25-year data veteran takes the helm; founder Kaycee Lai shifts to Chief Strategy Officer.

2025

Agentic platform & Mantra

Promethium launches its agentic platform and Mantra data-answer agent for self-service data at AI scale.

The Money

Backed for the long game

Early · 2019
$6M+
Series A · 2022
$26M

Total raised to date is approximately $34.5 million. The 2022 Series A was led by Insight Partners, with early backers .406 Ventures and Zetta Venture Partners - the latter an AI-first fund - continuing to participate. The company remains lean at around 34 employees, a deliberately small team selling into large, complex enterprise data environments.

Expertise

Who is behind it

CEO Prat Moghe brings 25 years in data and analytics. He founded and ran the cloud-analytics company Cazena, led teams at Cloudera and IBM Netezza, and holds a PhD in Electrical Engineering from UCLA. Founder Kaycee Lai, now Chief Strategy Officer, continues to drive product vision. The board and advisors include seasoned operators and investors such as Jocelyn Goldfein and Ganesh Bell.

The culture the company describes is collaboration-first, willing to challenge conventional approaches, and oriented toward bold, purpose-driven innovation. External markers back the ambition: Promethium has been named a Gartner Cool Vendor and holds SOC 2 Type II certification - table stakes for handling regulated enterprise data.

Off The Record

Details that amuse and inform

Questions & Answers

The short version

What does Promethium do?

It provides an Instant Data Fabric and agentic analytics platform that lets business users, analysts and AI agents get trusted, governed answers from data across many sources - without first moving or centralizing that data.

Who founded Promethium and who runs it now?

It was founded in 2018 by Kaycee Lai, who now serves as Chief Strategy Officer. Prat Moghe, a 25-year data-and-analytics veteran, is the current CEO.

How much funding has Promethium raised?

Roughly $34.5 million to date, including a $26 million Series A led by Insight Partners in 2022, with .406 Ventures and Zetta Venture Partners also participating.

What is Mantra?

Promethium's conversational data-answer agent. It turns natural-language questions into governed answers that include generated SQL, metadata, lineage and visualizations, and can publish those insights into existing tools and AI systems.

How is it different from a data warehouse?

Rather than requiring you to extract, transform and load data into a central warehouse, Promethium queries data where it already lives across 200+ cloud, on-premises and SaaS sources, adding context and governance so answers are trustworthy.