en.Wedoany.com Reported - Spanish artificial intelligence company Sherpa.ai has secured $18 million in funding to develop AI products that can be trained without ever accessing users' raw data. The company primarily serves clients such as banks, hospitals, and governments, who are most sensitive about where their data goes. Its technical approach aligns with the rapidly growing theme of "sovereign AI."

Sherpa.ai's core technology is federated learning. In this model, instead of aggregating data in one place for training, the model "goes to" the data. Each hospital or bank trains the model locally, sharing only learning experiences rather than raw records. Sherpa.ai claims its research can reduce the amount of data transmitted between sites by up to 99%. This approach is particularly important in regulated sectors with strict privacy laws, where data compliance requirements often hinder conventional AI projects. Founder and CEO Xabi Uribe-Etxebarria stated that the goal is to enable enterprises to "fully leverage the potential of AI without having to give up control, privacy, and sovereignty over their data."

Notably, Forgepoint Capital, a cybersecurity and AI investor from Silicon Valley in the United States, participated in this funding round, along with existing investors Mundi Ventures, Ekarpen, Allegra Holdings, and SETT. For a company built on the anxiety of keeping AI data within Europe, recognition from the U.S. is significant. Sherpa.ai's client list also points in the same direction. The company has recently signed contracts with Spain's Indra, banks Caja Laboral and Unicaja, security group Prosegur, genomics company Centogene, and the US National Institutes of Health (NIH). A privacy-focused European company selling products to a U.S. federal agency is a testament in itself.

"Sovereign AI" is a crowded label, and federated learning is not a new concept. However, Sherpa.ai's version carries weight due to its solid research foundation. The company has published peer-reviewed papers on training large language models across private datasets and has collaborated with the US National Institutes of Health (NIH) and University College London to apply this technology to rare disease diagnosis. This funding round is moderate in size, and the field is highly competitive. From national model projects to privacy-focused security startups, many companies promise to develop AI that respects data boundaries. Yet the market demand is real and growing. As governments impose stricter regulations on where data is stored, solutions from companies that can train models without ever touching data will become more compelling. Sherpa.ai's $18 million funding round is a bet on this future.