← PortfolioCompute and infrastructure

Qernelzoo

Quantum-resistant cryptography accelerator

Team
Oluseun Omonijefounder
Founded
2026
Invested
2026
Links
The problem

How do you change every lock on the internet before someone can pick them?

Click anywhere to edit the one policy, and watch every service follow at once. Point at a service to find its line in the bill of materials.A scan finds every lock across a dozen services and lists each one in a bill of materials. Then every call site is rewritten to ask one policy for its keys, rather than naming a scheme itself. After that, moving to post-quantum schemes, or to a hybrid pairing, is a single edit that reaches every service at once.An illustration, not real data.
How the locks work

Most of the that guards the internet rests on one of three maths problems: factoring big integers, the discrete logarithm, and its elliptic-curve cousin. RSA is safe only because factoring large numbers is out of reach, and no ordinary computer is known to factor integers in polynomial time.

changes that. On a quantum computer it factors in polynomial time, and it could break RSA and both flavours of Diffie–Hellman key exchange. The catch, for now, is size. Beating a classical machine may take millions of , because quantum error correction eats so many of them.

So cryptographers built new locks. NIST has finalised standards designed to withstand a quantum attacker: ML-KEM, formerly Kyber, for setting up encryption keys, and ML-DSA, formerly Dilithium, for digital signatures. Both are . Their security leans on the shortest vector problem, which is thought to be hard to solve efficiently even with a quantum computer.

Further reading Post-quantum cryptography (Wikipedia)Shor's algorithm (Wikipedia)NIST Releases First 3 Finalized Post-Quantum Encryption Standards (NIST)Lattice-based cryptography (Wikipedia)

Why it is hard
  1. i.

    Steal now, read later

    Nobody needs a quantum computer today to plan for one. "Harvest now, decrypt later" means copying encrypted traffic you can't read yet and storing it until decryption catches up. Plenty of data recorded now will still be sensitive for decades, so the clock on a migration starts well before the machine exists.

  2. ii.

    Heavier keys

    The new locks are bulkier. Many post-quantum algorithms need larger keys than the ones they replace. At its middle security level, an ML-KEM public key is 1,184 bytes and a 1,088. In one chat-encryption test, swapping the elliptic-curve exchange for ML-KEM made it about 2.3 times slower and added roughly 70 times more data overhead. Most of that runtime goes on internal hashing, which is the kind of work that benefits from hardware acceleration.

  3. iii.

    Crypto hides everywhere

    Cryptography sits in TLS connections, SSH keys, code signing, certificate authorities, and the firmware of hardware security modules and IoT devices. When every product picks its own scheme in code, each swap turns into a hunt across the network. The standards body's own advice is to start now, because full integration will take time. It has happened before: DES, 512-bit RSA and RC4 were all once considered secure.

Further reading Harvest now, decrypt later (Wikipedia)Post-quantum cryptography (Wikipedia)Kyber (Wikipedia)Quill: Crypto-agile security against quantum attacks (Qernel Technologies)NIST Releases First 3 Finalized Post-Quantum Encryption Standards (NIST)Cryptographic agility (Wikipedia)

What Qernelzoo is after

Qernelzoo works on efficient cryptography, with an eye on the day fault-tolerant quantum computers arrive. Its first product, Quill, is about the unglamorous middle of the problem: finding the old locks, working out which to change first, and making sure the next change is cheaper than this one.

Further reading Qernel (Qernel Technologies)Quill: Crypto-agile security against quantum attacks (Qernel Technologies)

How they go at it
  1. Step 1: Find every lock

    Quill scans code and networks, writes a cryptographic bill of materials, or CBOM, and turns each finding into a prioritised path to post-quantum. It can also grade a domain from the outside, reading the TLS key exchange, certificate chains and signature algorithms it already publishes.

  2. Step 2: One place to change

    The bigger idea is : being able to swap cryptographic algorithms with ease and at least partly automatically. Instead of every service picking its own scheme in code, call sites resolve their keys from one policy, so the migration after this one is a single edit.

Further reading Quill: Crypto-agile security against quantum attacks (Qernel Technologies)Cryptographic agility (Wikipedia)

Still open
  • When will a quantum computer big enough actually show up?

    Nobody knows. Lab demonstrations of Shor's algorithm have only factored small numbers, and one team's machine managed 15 and 21 but not 35. Even so, some experts predict a code-breaking device within a decade.

  • How much should we trust the new maths?

    Lattice schemes like NTRU have been studied for many years without a feasible attack. Another widely noticed post-quantum scheme, SIDH/SIKE, was spectacularly broken, though that attack only works on its own family. That's why some deployments use , pairing each new post-quantum scheme with a more proven classical one.

Further reading Shor's algorithm (Wikipedia)Post-quantum cryptography (Wikipedia)NIST Releases First 3 Finalized Post-Quantum Encryption Standards (NIST)

About Qernelzoo

Qernelzoo is a seed-stage cybersecurity company developing quantum-resistant cryptography acceleration software.

Words used here
public-key cryptography
Encryption where anyone can lock a message with a public key but only the holder of the matching private key can unlock it.
Shor's algorithm
A quantum algorithm that factors integers and solves discrete logarithms fast enough to break today's public-key systems.
qubits
The basic units of a quantum computer, which unlike ordinary bits can hold a mix of 0 and 1.
lattice-based
Built on geometric grids of points in many dimensions, where finding the shortest step between points is believed to be very hard.
ciphertext
The scrambled output of encryption, which is what actually travels over the wire.
crypto-agility
Designing systems so their cryptographic algorithms can be swapped without rebuilding everything around them.
hybrid encryption
Running a new post-quantum scheme alongside a proven classical one, so data stays safe if either holds.
Sources