Nera Systems Brings Confidential AI to Google Cloud
Regulated organizations can now use Gemini on sensitive data without exposing it to the model
PALO ALTO, Calif. — August 13, 2026 — Nera Systems today announced that its confidential AI platform is available on Google Cloud, allowing organizations in regulated industries to analyze their most sensitive data using frontier AI models without that data ever being exposed.
Most enterprises share the same problem. The data that would benefit most from AI is the data they are least able to use. Client financials sit behind confidentiality obligations. Patient and customer records sit behind HIPAA, GLBA, and GDPR. Proprietary datasets sit behind contracts that never anticipated an AI model reading them. The work that would create the most value is the work that never gets approved, while staff quietly turn to consumer AI tools with data that should never leave the building.
The pressure is increasing from several directions. Regulators have begun asking organizations what their AI systems can access. Insurers have started writing AI exclusions into corporate policies. The usual answers, a written policy or a promise from a vendor, are becoming harder to defend.
Nera addresses this at the architectural level. Data is encrypted before it leaves the customer's control, using keys the customer holds. The AI model receives the user's question and the structure of the data, never the underlying values. Computation runs on encrypted data, and results are decrypted only by the customer. Nera itself cannot read the data at any point.
The result is that the exposure question stops depending on trust. No plaintext copy of the customer's data sits anywhere a breach, a policy change, or a vendor incident could reach.
Running on Google Cloud, Nera customers can put Gemini to work on data that was previously off limits: cash flow and variance analysis on client financials, anomaly detection across sensitive operational records, and risk analysis on data that could not be shared with a third-party model.
"Every enterprise wants AI to generate value from all of their data, but the most valuable data is often the confidential data they can't afford to expose. Nera removes that tradeoff. Data stays encrypted even while the AI is working on it, so customers can safely use tools like Gemini on information they would never otherwise risk exposing," said Rami Akeela, Ph.D., Founder and CEO of Nera Systems. "The teams that need this most have usually been told to stay away from AI entirely. We want to bring that data back into use, safely."
"Confidential computing and generative AI are converging, and enterprises in regulated industries need both. Programs like Startup Success exist to help companies working on that problem build on Google Cloud." — James Tromans, Senior Director of Software Engineering, Google Cloud
Nera is designed around three outcomes for regulated teams:
Meet compliance requirements without slowing down. Use AI on data governed by client confidentiality obligations, HIPAA, GLBA, or GDPR without waiting on a policy exception for every project.
Reduce data exposure risk. Because data is never decrypted, including during AI processing, there is no plaintext copy for a breach or vendor incident to reach.
Keep the models teams already want. Nera works with leading commercial models, so organizations are not tied to a single vendor.
For example, an accounting firm that currently prohibits staff from entering client financial data into AI tools can use Nera to analyze that same data, running variance analysis or flagging anomalies, without the data being exposed in plaintext and without conflicting with client confidentiality obligations.
Nera Systems came up through Google Cloud's Startup Success Program, and the two companies are continuing that work together, including joint engagement with enterprise customers in regulated industries.
About Nera Systems
Nera Systems is a confidential AI company that enables organizations to use their most sensitive data with commercial AI systems, without ever exposing it. Nera's platform keeps data encrypted throughout its lifecycle, including during processing, so that no third party, including Nera itself, can access it in plaintext. Its flagship product, Nera ChatApp, lets teams query sensitive structured and numerical data using leading LLMs including Gemini, Claude, and ChatGPT. Nera is built for organizations that handle data too sensitive to put into standard AI tools, including accounting firms, financial institutions, insurers, and enterprise finance teams. Nera is a Techstars alum and a member of the Google Cloud Startup Success Program. Learn more at nera.systems.
Media contact: Nera Systems — info@nera.systems