The flight recorder for enterprise AI. Every prompt, every response, every configuration - captured, monitored, and made audit-ready inside your own cloud.
The mark: six nodes wired to one core. A picture of the pitch - many models, tools and teams, all reporting to a single system of record.
Somewhere inside a large enterprise right now, a generative AI application is drafting an email, approving a refund, or summarizing a contract. It is fast. It is helpful. And when the auditor, the regulator, or the general counsel asks what exactly it told the customer - and whether it was allowed to - the honest answer, for most companies, is a shrug.
SolidCore.ai exists to replace the shrug with a record. The Menlo Park startup builds the security and governance layer that sits beneath first-party enterprise GenAI: it captures every prompt, response, configuration, and scrap of metadata, stores it immutably inside the customer's own cloud, and watches the whole thing for misconfigurations, policy violations, and usage that looks off. When the hard questions come, the receipts are already filed.
It is a decidedly unglamorous idea in an industry drunk on demos. Nobody tweets a screenshot of a compliance log. But that is rather the point - the interesting companies are usually the ones handling the parts everyone else would prefer to skip.
Prompts, responses, configuration data and metadata from every GenAI interaction - stored inside your own cloud, giving full transparency across teams, tools and environments. No proxy in the data path.
Real-time detection of misconfigurations, access violations and anomalous usage patterns, with alerting and recommended resolution workflows so problems surface before they become incidents.
Automated compliance monitoring and defensible, audit-ready evidence aligned with NIST AI RMF, ISO/IEC 42001 and the EU AI Act - the alphabet soup turned into a green checkmark.
SolidCore gives enterprises the missing infrastructure layer they need to move with speed.
Second act in infrastructure security. Chiu founded HyTrust - which secured cloud and container infrastructure for demanding enterprises and was acquired by Entrust. A UC Berkeley materials science graduate who traded materials for the harder problem: securing the thing that talks back.
Former CTO for M365 Security and Compliance at Microsoft, then Microsoft Security's AI Security, Privacy and Data Residency Architect. A contributor to NIST and Cloud Security Alliance standards - which is to say, she helped write some of the rules SolidCore now helps enterprises pass.
The stakes are incredibly high for every organization to integrate GenAI into their existing products.
Rivals who rarely agree on anything each backed SolidCore at seed stage - a rare signal about where enterprise AI governance is heading.
Eric Chiu and Hemma Prafullchandra found SolidCore.ai in Menlo Park, betting that AI governance is the next infrastructure layer - not a feature bolted on later.
The company emerges from stealth with $4M in seed funding led by Runtime Ventures, and reveals partnerships with all four major AI infrastructure providers.
Coverage spreads across FinSMEs, The AI Insider, Funded.com and others, framing SolidCore as an early answer to the enterprise GenAI trust gap.
The AI still drafts the email. It still approves the refund. It is still fast. But now, when the auditor asks what it said and whether it was allowed to, nobody shrugs. There is a timestamped, immutable line in the log - and a monitor that already flagged the one time it drifted.
SolidCore's wager is simple: enterprises will trust AI the same way they learned to trust the cloud - not on faith, but on visibility. Trust, it turns out, is not a slide in a deck. It is a log line with a timestamp. That is the unglamorous, necessary thing SolidCore is building.