Pomegra
Nashville startup Enigmata closed a $6.5M seed round led by Blockchange Ventures on September 10, 2026, commercializing a cryptographic platform that lets AI models train and run inference without ever touching plaintext.
- Blockchange Ventures led the seed; valuation undisclosed
- Enigmata Cipher matches raw-data model accuracy while cutting training time 8%-10%, per internal benchmarks
- Target markets include banking, healthcare, life sciences, insurance, and data licensing
What Did Enigmata Actually Build?
Enigmata Cipher is a patent-pending cryptographic layer that converts records, documents, and datasets into an encrypted form that AI systems - training pipelines, search tools, and analytics engines - can process directly on existing enterprise hardware. The underlying data never appears in plaintext. Not to the model. Not to the infrastructure running it.
That is the crux of the problem Enigmata is trying to solve. When a health system wants to train a diagnostic model, or a bank wants to run a fraud-detection pipeline, the data involved is among the most legally and competitively sensitive material the organization holds. Existing approaches require decrypting data before it enters an AI workflow, creating a window of exposure that privacy laws, insurance regulators, and internal security teams increasingly treat as unacceptable.
Cipher routes around that exposure entirely, doing the computational work on the encrypted form. CEO and co-founder Scott Searle has framed the long-term ambition as infrastructure for a secure data economy - one where institutions can license encrypted datasets for AI training under enforceable usage terms and retain control of the underlying assets.
Why Does the Funding Amount Matter Here?
The $6.5M figure (the prompt referenced $4.5M; verified sources confirm $6.5M) is modest for a company tackling what is genuinely a hard problem in applied cryptography. Fully homomorphic encryption - the theoretical ceiling of what Enigmata is approaching - has historically carried computational costs that made production deployment impractical. The performance claims Enigmata is making, an 8%-10% improvement in training speed on encrypted data relative to raw data, are significant if they hold up outside internal benchmarks. Independent reproduction of those numbers will determine how seriously the enterprise market takes the platform.
Blockchange Ventures, the lead, backs companies at the intersection of cryptography, data infrastructure, and financial services - a thesis that maps cleanly to Enigmata's target buyers. No other investors were disclosed.
Who Is the Customer?
The design target is explicit: banks, insurers, health systems, life-science organizations, publishers, and data providers. Each sector holds data too sensitive to expose to third-party AI systems under standard conditions. Each is also under competitive pressure to use AI for exactly the workflows that would require exposing that data.
Healthcare is the most obvious case. Patient records are protected by a web of federal and state law; the consequences of a breach are severe. But the same model that diagnoses disease more accurately needs to see patient records to learn. Enigmata Cipher is positioned as the mechanism that makes training on protected health information viable without special regulatory carve-outs or expensive secure-enclave infrastructure.
Publishing and data licensing add a second revenue hypothesis. If a media company or data broker can make encrypted datasets available for AI training while retaining verifiable control of the underlying content, the commercial model for licensed data changes. The question is whether buyers in those markets will trust the enforceable-usage claim enough to act on it.
What Is Enigmata Not?
It is not a differential-privacy tool, which adds statistical noise to outputs. It is not a secure-enclave approach, which isolates computation inside hardened hardware. And it is not synthetic-data generation, which sidesteps the original data altogether. Cipher claims to let AI systems work on the real data, in encrypted form, without performance degradation. That is a narrower and more technically demanding claim than any of the alternatives - which is both the risk and the pitch.
The company also says Cipher supports targeted record deletion without full model retraining. That capability, if it works at scale, addresses a specific compliance requirement that has proven expensive for organizations subject to right-to-erasure rules.
Outlook
Enigmata emerges from stealth with a coherent problem statement, credible target markets, and benchmark data that needs external validation. The $6.5M seed is enough to drive commercialization; it is not enough to win a market. The next proof point is whether the accuracy and speed claims survive contact with production data at a real enterprise customer. If they do, the data-licensing infrastructure play becomes the more interesting story. If they do not, the company has a technically sophisticated answer to a question the market has not yet agreed to ask.
