AI isotopes He left Stealth on Thursday with a healthy round price $ 20 million.
He offers an AI agent to solve the problem that data evaluation products have been faced with for many years: individuals who know the right way to start large data sets infrastructure are not those that actually have to make use of data.
In the case of LLM, business managers can ask questions about their data in natural language. Agent Isotopes, Aidnn, can answer and develop complex planning documents, collect data from the storage site, reminiscent of financial applications, ERP, CRM and cloud churning.
There are countless offers of agency business evaluation, but the co -founders of isotope have a unique pedigree. Their product is so sophisticated that the startup has already submitted an application for 10 patents, said co -founder Arun Murty.
Slightly over 20 years ago, when Murty was in the mid -Nineteen Twenties, he worked in Yahoo in a team that built an open source project called Hadoop. Hadoop stimulated the initial madness of enormous data sets from 2010.
In 2011, Yahoo turned him into a company called Hortonworks, and Murty as a co -founder and product director. Just 4 years after starting Hortonworks made public. But the growth of latest storage technologies in the cloud was reflected on the Hadoop market, and Hortonworks eventually merged with the largest rival, Clouder. The combined company was private in 2021 after the famous activist investor Carl Icahn was involved.
Murty went to Clouder for several years during this confusion, managing about 200 people. However, he says that even there he saw the everlasting problem with access to data. He remembers quarterly conference connections with analysts from Wall Street Grilling Execs regarding the details of the motion.
TechCrunch event
San Francisco
|.
October 27-29 2025
They couldn’t all the time answer because they didn’t have access to data. “It was embarrassing,” he admits now. “We were a company of large data sets.”
In 2021 he left this work without a real plan, what to do next, and VC introduced him to the scaling of AI Alexander Wang. After a few conversations with Wang and a little bit of consulting work, Murty joined Scale as a technology director.
It was like “obtaining a large -scale doctor”, he describes. “Understanding what drives these models and how to improve them.”
But when his old buddy from Hortonworks, he called him Prasanth Jayachandra, they each decided to make his own startup AI. They convinced their third asshole, Gopala Vijayaraghavana, also from the time of Hortonworks, to hitch them, and at the end of 2024 they founded isotopes. Their seed round was led by Vab Goel in NTTVC. (Goel was previously in Norwest Ventures.)
The background of the founders allowed them to build an agent who can find data from the storage site (whether Salesforce or snowflake), but also clean the data. The agent also maintains a lot of contextual memory as useful for complex tasks.
“It’s far beyond a simple chatbot,” said Murty. For example, if the team asks Aidnn to develop a report on monthly repetitive revenue trends: “The data you want to talk to do not exist, at least a form with which you need to talk. This is a multi -stage plan, a very complex plan: separation of metadata, reading data, cleaning and normalizing, attach to data, proprat revenue, aggregate.”
The agent also shows his steps, reasoning, assumptions and suggestions for anomalies in the data. There will even be recommendations on the right way to proceed. Isotopes also guarantees that corporate customers can implement without sharing any data to the manufacturers of AI models supplying an agent.
Despite this, although sophisticated, like isotopes, the startup is in the face of high competition. The current ones like Tableau Salesforce Already offer agents (among the primary agent AI Salesforce), and on the market there are many other founders of startups with impressive pedigrees reminiscent of Wisdomai.