Quside, spun out of Barcelona's ICFO photonics research institute, is shipping the Ruby N1 -- an SP800-90B certified quantum random number generator chip that uses photonic quantum effects rather than electronic noise or oscillator jitter to produce verifiable entropy. The company claims it has shipped to more than 50 customers across defense, space, HPC, and data centers and expects profitability within 24 months.
The timing is not accidental. Two forces are converging: the industry-wide migration to post-quantum cryptography (PQC), which in hybrid mode runs classical and PQC algorithms in parallel and significantly increases key-generation and certificate workloads; and the expansion of agentic AI infrastructure, where software agents transacting on behalf of users generate far more cryptographic handshakes per second than human-driven sessions ever did. Both put pressure on the entropy source quality and throughput that conventional pseudo-random and hardware-noise-based generators were never designed for.
The differentiator worth examining is verifiability. Classical TRNG approaches generate randomness but give you limited ability to certify the physical origin of that entropy in operation. Quside's photonic approach is grounded in quantum optics that can be modeled from first principles, which is how they achieve the SP800-90B certification with an auditable physical basis rather than just statistical test-passing. For any system that needs to pass a FIPS audit for cryptographic module validation, that distinction is real.
The caveat: Quside is still a startup. The Ruby N1 addresses a real problem, but integration into ASIC designs and HSM architectures is rarely straightforward, and commercial traction at 50+ customers is solid but not yet proven at hyperscale volumes. Watch whether they close a Tier 1 data center design-win in the next 12 months -- that would validate the market timing thesis.