Q.ANT photonic computing processor module.
Supporting photograph for this technology report.

German photonic-computing company Q.ANT has unveiled a new processor designed to perform selected mathematical operations using light rather than relying entirely on conventional electronic calculations. The company introduced its Native Processing Unit Gen 2+ at the SC26 supercomputing event in Chicago, saying the processor adds optical Discrete Fourier Transform capabilities to its existing photonic architecture. Q.ANT’s November 17 announcement describes the technology as a step towards using photonic chips as specialised co-processors alongside traditional CPUs and GPUs.

The central idea is to use the physical properties of light to carry out certain computations in a single optical operation. Fourier transforms are mathematical tools used in signal processing, image analysis, simulations and other data-intensive applications. Q.ANT says performing these operations directly in light could reduce the amount of work required from electronic processors for selected workloads. The company argues that this may improve efficiency in applications where large volumes of data must be transformed repeatedly. The technical description published by Q.ANT explains how its NPU Gen 2+ expands the set of functions available through photonic computing.

At SC26, Q.ANT demonstrated a hybrid system that paired its photonic processor with an AMD GPU to run a large language model using optical Fourier transforms. The company said the optical processor handled a mathematical operation associated with the model’s attention layer, a computationally demanding part of many modern language models. The demonstration is significant because it suggests photonic hardware could complement conventional accelerators rather than attempting to replace them completely. The company’s demonstration announcement identifies the AMD GPU and optical NPU as components of the hybrid system.

Photonic computing has attracted interest because data centres face growing electricity and cooling demands as AI workloads expand. Conventional chips process information through electronic signals moving through transistors and circuits, while photonic systems use light to perform selected operations. That does not mean every computing task can immediately be transferred to optical hardware. Memory, data conversion, control systems and software integration remain important engineering challenges, and a photonic accelerator is likely to be most useful when it handles operations suited to its architecture. Q.ANT presents the NPU as a complementary processor within hybrid systems, not a universal substitute for GPUs. Its product announcement sets out that intended role.

The company says it plans to offer commercial systems combining its Gen 2 and Gen 2+ processors in the second half of 2027. Independent performance comparisons, large-scale deployments and real-world energy measurements will be important in determining how quickly the technology can move beyond demonstrations. If photonic processors can deliver measurable gains on suitable workloads, they could become part of a more diverse AI hardware ecosystem. For now, Q.ANT’s announcement is an encouraging research and product milestone, while the commercial impact remains to be demonstrated. Q.ANT’s roadmap and technical details provide the company’s current outlook.