Back to home

White Papers

Technical deep dives into the cryptography, architecture, and threat models behind Novaqore AI.

Post-Quantum Key Encapsulation in Practice

How Novaqore AI uses CRYSTALS-Kyber

An overview of how we implement Kyber-1024 for per-request key encapsulation, why we chose it over other post-quantum candidates, and how it integrates with AES-256-GCM. Includes the full encryption lifecycle from client to server and back.

Coming soon

Why HTTPS Is Not Enough for AI

Attack surfaces beyond TLS

A detailed look at the real-world risks that exist even when TLS is in place. Covers TLS inspection by corporate proxies, certificate authority compromise, CDN termination, provider-side plaintext access, and the harvest-now-decrypt-later quantum threat.

Coming soon

Zero-Knowledge AI Infrastructure

How we built it so even we can't see your data

How the architecture is designed so that even Novaqore cannot access your data. Covers zero-logging, in-memory-only decryption, per-request key generation, and why we don't store shared secrets or session data.

Coming soon

Threat Modeling for Private LLMs

What we defend against and how

A walkthrough of the threat model: man-in-the-middle attacks, insider threats at the infrastructure level, supply chain attacks on dependencies, and quantum adversaries. Explains how each layer of the design addresses these threats.

Coming soon

The Harvest Now, Decrypt Later Problem

How post-quantum cryptography solves it

A focused paper on the quantum threat to today's encrypted communications. Why RSA and elliptic curve cryptography won't survive quantum computing, how Kyber's lattice-based approach resists quantum attacks, and what this means for AI data sent today.

Coming soon

Want to know when these are published?

Follow us on X / Twitter or join Discord for updates.

© 2026 Novaqore AI. All rights reserved.