Blind Insight

Blind Insight builds encryption-in-use technology that lets organizations query, analyze, and run machine learning against sensitive data without ever decrypting it. Its core product, Blind(L)LM, works as a semantic encryption proxy: data is encrypted at the source before it leaves the customer’s own infrastructure, computation runs directly against the ciphertext, and results come back still encrypted, decryptable only where the customer holds the keys. That is a meaningfully harder technical problem than typical encryption-at-rest or in-transit approaches, since most systems have to decrypt data before they can compute or search over it.

Founded in 2022 and based in Marina del Rey, California, Blind Insight is a very small company, around four employees, that has raised a pre-seed round of roughly $1.15 million from Responsibly Ventures, Women’s Equity Lab, Techstars, 500 Global, and Altari Ventures. The company targets regulated industries such as healthcare and financial services that need to apply AI and analytics to sensitive data but face legal or contractual barriers to broader data access, and it publishes a specific performance claim (90 aggregate queries across 1 million encrypted records completed in 58 seconds) as evidence its approach is usable in production rather than purely theoretical.

The company positions its platform against DORA, GDPR, HIPAA, SOC 2, and FIPS compliance requirements, and offers both a hosted option (Blind Insight manages infrastructure while the customer retains encryption keys) and a self-hosted Docker deployment. Given the company’s early stage and small size, its performance and compliance claims should be treated as vendor-published rather than independently benchmarked at this point.