Q.ANT and IONOS launch photonic computing pilot

Q.ANT and IONOS have launched a pilot programme allowing selected developers to execute computational workloads on physical photonic processors hosted within German data centre infrastructure.

The arrangement gives approved participants access to Q.ANT's second-generation Native Processing Unit through dedicated IONOS bare-metal servers. Developers can submit their own software and data for evaluation, moving beyond the simulated environment provided by Q.ANT's open-source development kit.

Participation is limited and subject to an application process. The programme begins in November 2026, with accepted organisations receiving technical onboarding, project support, and assistance evaluating their results. Q.ANT and IONOS have not established a timetable for unrestricted commercial availability.

The photonic processor performs selected mathematical operations using light travelling through integrated optical circuits. Its architecture differs from conventional digital processors, which use electronic transistor circuits to execute arithmetic, logic, and data movement operations.

Q.ANT's hardware uses photonic integrated circuits fabricated with thin-film lithium niobate. This material has electro-optic properties that allow electrical signals to influence the behaviour of light passing through appropriately designed optical structures.

The processor is intended to exploit the physical properties of these circuits to perform mathematical functions directly within the optical domain. Q.ANT identifies matrix multiplication and nonlinear operations as suitable workloads, particularly when algorithms can be structured around the characteristics of the hardware.

Nonlinear mathematical functions are used in artificial intelligence and scientific computing to represent relationships that cannot be expressed through a simple linear transformation. In conventional digital systems, the required calculations may be implemented through sequences of arithmetic operations, approximations, or stored function values.

Q.ANT is developing an alternative approach in which certain functions are implemented through the response of an optical circuit. Such an operation may reduce the number of digital processing steps required for a suitable workload, although the resulting advantage depends on the algorithm, the hardware configuration, and the interfaces connecting the photonic processor to the rest of the computer.

Optical computation does not eliminate the need for conventional electronics. Input information must be prepared and supplied in a form compatible with the photonic circuit, while the resulting optical output requires measurement and processing before it can be used by other software or hardware.

These conversion and interfacing stages contribute to the total energy consumption and execution time. A reduction in the cost of an individual optical operation does not automatically establish a corresponding improvement for the complete computing system.

The suitability of a photonic processor also depends on numerical accuracy. Different applications tolerate different levels of variation in calculated results, and the behaviour of an optical device may be affected by operating conditions, component characteristics, and the precision of associated electronic interfaces.

An image-processing model may accommodate some numerical differences without materially changing its output, while a scientific calculation could require tighter limits. Establishing acceptable performance therefore involves examining the behaviour of a complete workload rather than relying solely on the nominal capabilities of the processing hardware.

The new pilot provides an opportunity to collect these measurements using physical processors. Participants will run selected applications through the IONOS environment, allowing the results to be compared with existing implementations or development expectations.

Q.ANT identifies industrial image processing, time-series prediction, scientific simulation, and robotics as possible application areas. The programme is also open to other computational workloads compatible with the processor architecture.

Applicants must provide a concrete use case and have the technical capability to administer a bare-metal server independently. Unlike a conventional managed cloud application, the environment requires participants to work directly with a dedicated computing configuration and its supporting software.

The pilot is free for a limited number of selected participants. Q.ANT will provide technical onboarding, project support, and a review of the workload results, followed by a workshop to assess the performance information obtained during the evaluation.

IONOS is responsible for the underlying server infrastructure, with the photonic hardware operating in a German data centre. The arrangement places the hosted workloads under German jurisdiction and provides European data residency for the relevant infrastructure.

The physical location may be important where organisations restrict the geographical location of their computing resources. Data residency, however, is only one aspect of information security and does not independently establish compliance with every regulatory or application-specific requirement.

The programme builds on Q.ANT's September 2026 release of its open-source software development kit. That software allowed developers to investigate the processor architecture and simulate applications before they had access to deployed hardware.

Moving from simulation to physical execution will expose characteristics that cannot be established solely through software models. Processing latency, numerical behaviour, hardware interfacing, and the energy consumed by the complete system can be measured under the conditions of the hosted installation.

Q.ANT's second-generation processors have also been deployed in specialist research environments, including the Jülich Supercomputing Centre. The IONOS arrangement introduces an additional access route through a commercial infrastructure provider, allowing selected external developers to assess the hardware without installing a photonic processor at their own premises.

The processor is manufactured using thin-film lithium niobate photonic technology. Q.ANT operates a pilot production line for the material platform in Stuttgart, supporting development of the optical circuits and their integration into computing systems.

Scaling the architecture beyond evaluation projects will require further assessment of manufacturing repeatability, system integration, software compatibility, and the economics of operating photonic hardware alongside established computing resources.

Q.ANT and IONOS will use the results obtained by pilot participants to inform subsequent development of the processor and its software stack. The companies have not published independent system-level benchmarks establishing universal performance or energy-efficiency advantages for the new hosted configuration.