A Dutch startup just closed a photonic control loop in 150 nanoseconds—faster than light travels 30 meters of fiber.
The loom's 32 threads and shuttles mirror the 32-input, 32-output routing matrix of QuiX Quantum's control unit.
QuiX Quantum, based in Enschede, Netherlands, announced the first installation of its Feed-Forward Control Unit (FFCU), the hardware that converts single-photon measurement signals into control actions on photonic chips in real time. The rack-mounted system delivers a deterministic 150-nanosecond execution loop from detector input to settled output voltage on the chip. In that same interval, light in standard telecom fiber covers roughly 30 meters. The control loop is now faster than the physical transit of the photon carrying the information. Photonic quantum computing has its first real-time, closed-loop nervous system, and the latency race across all quantum modalities now has a public benchmark.
The Missing Piece for Universal Photonic Quantum Computing
This new precision mapping symbolizes the shift from speculative quantum computing to deterministic, measurable progress.
Measurement-based quantum computing runs on a sequence of measurements. One measurement’s outcome determines which operation comes next. Without a control loop that can read a photon detector, compute a decision, and apply a phase shift to the next operation before the next photon arrives, the system cannot adapt. It cannot reach universality. Feed-forward control is the component that closes that loop. Until now, it was a lab patchwork. QuiX Quantum made it a product.
The FFCU is one module in a full-stack architecture that spans photon generation, multiplexing, state generation, measurement, and photonic assembly control. In May 2026, the company introduced the Photonic Assembly Control Unit (PACU), a rack-mountable system hosting photonic chips with up to 1,000 low-speed phase shifters and 32 high-speed phase shifters. The FFCU connects to that control plane, creating a complete signal chain from photon detection to chip reconfiguration. The German Aerospace Center (DLR) has put €14 million behind the architecture, with a contract that specifies delivery of 8-qubit and 64-qubit fully universal photonic quantum computers.
What 150 Nanoseconds Actually Means
The FFCU hardware is a dual-FPGA processing bus with a custom analog front-end. Two FPGA blocks share a high-speed, low-latency communication bus and feed a routing matrix with 32 dedicated signal inputs and 32 settled voltage outputs. The system runs a fixed 150 ns latency window. Not best-effort. Not average. Deterministic.
Determinism matters because quantum error correction and measurement-based algorithms require operations to happen inside a predictable coherence window. A control system that sometimes delivers a decision in 200 ns and sometimes in 400 ns introduces timing jitter that compounds across thousands of gate operations. An academic paper on FPGA-based feedforward for photonic systems, available on the arxiv preprint server, reports a total system latency of 196 ns. QuiX’s 150 ns is 23 percent faster and, critically, a guaranteed ceiling rather than a measured run.
“Universal photonic quantum computing requires more than high-quality photonic chips. It requires a complete system stack that can generate, route, measure and control photons in real time,” said Stefan Hengesbach, CEO of QuiX Quantum. Andrew Roos, VP of R&D, framed the speed benchmark: “To put that timing in perspective, in 150 nanoseconds light travels only about 30 meters in telecom fibre.”
The Consensus Is Treating This as a Photonic Quantum Computing Story. It’s Not.
The FFCU is a classical computing achievement with quantum implications. The 150 ns latency comes from FPGA logic and analog circuit design, not from any quantum phenomenon. The uncomfortable implication is that the near-term viability of any quantum modality now depends less on qubit quality and more on the speed and determinism of the classical logic that controls the quantum layer. Superconducting qubits, trapped ions, neutral atoms—all require measurement, decision, and reconfiguration inside a finite coherence budget. Every modality has a control stack. Almost no one publishes its latency numbers. QuiX Quantum just did.
The industry obsesses over quantum volume, qubit counts, and coherence times. Those metrics describe potential. Control-loop latency describes capability. A system with a million high-coherence qubits and a sluggish, jittery control stack is a physics demonstration. A system with modest qubit counts and a deterministic sub-200 ns control loop is a computer. The distinction is now explicit, and the other modalities have no comparable public benchmark to point to. Their silence on feed-forward latency is itself a data point.
Here is my prediction: Within 18 months, QuiX Quantum will demonstrate a 64-qubit photonic processor executing a non-trivial, measurement-based algorithm that a classical computer cannot simulate in real time. That demonstration will force superconducting and trapped-ion competitors to disclose their own feed-forward latency benchmarks. The DLR will double down on photonics as its primary path to utility-scale quantum computing. At least one major U.S. quantum company will acquire or partner with a photonic control startup to avoid falling behind in a race it did not realize had started.
The Metric That Matters Is Nanoseconds, Not Qubits
If you are mapping the quantum landscape, stop sorting companies by qubit count. Start demanding latency benchmarks for the control stack. Ask what the worst-case, deterministic control-loop latency is. Ask whether it scales linearly or degrades as qubit counts rise. The companies that can close the loop fastest and most predictably will be the first to unlock practical quantum advantage. The metric that separates a lab experiment from an engineered system is not quantum volume. It is nanoseconds from detection to action.
In the time it took to read this article, a photon in QuiX’s lab completed thousands of feed-forward cycles. The era of photonic quantum computing as a real-time systems discipline has arrived. The physics was always clever. The engineering just got faster than the speed of light in glass.
