Two Talpiot alumni and Israel Defense Prize winners raise $10 million for Upriver
Upriver has raised $10 million, led by Valley Capital Partners, which has invested in companies such as Decart and Factify, and Hetz Ventures, which invested about $4 million in the early round. The round also included prominent angels, among them Lou Sirer, founder of New Relic; Yotam Segev and Tamar Bar-Ilan, founders of Cyera; and Abe Gong, founder of Great Expectations.
Upriver was founded in 2024 by Ido Bronstein and Omri Lifshitz, graduates of the same Talpiot class and holders of master's degrees in AI. After completing the program, Bronstein served for six years in Unit 8200, where he managed cyber operations and later worked on data platforms, and he won the Israel Defense Prize. Lifshitz managed R&D in cyber at the Prime Minister’s Office, where he led the development of complex data systems designed to collect, process and make large volumes of information accessible in real time. He also won the Israel Defense Prize for his work. After completing their service, the two decided to found a company together, but October 7 interrupted their plans and they were both called up for an extended and intensive reserve service, after which they established the company.
Upriver employs 21 people in Israel and the United States and is already working with organizations such as Unity and DMGT, as well as with leading data platforms including Databricks and Snowflake. Upriver has developed an AI agent platform that autonomously manages the enterprise data layer. The system connects to the organization’s data environment, identifies and fixes data quality issues, maintains existing data processes and creates new databases. It is also able to map the entire enterprise data environment using a dedicated Context engine, and operate AI agents that work together and can analyze and verify information even in complex data environments.
Ido Bronstein, co-founder and CEO of Upriver, said, "We see organizations investing heavily in AI, but struggling to generate real impact from it because their data simply is not ready."
