STUTTGART, Germany and CHICAGO, Nov. 17, 2026 (GLOBE NEWSWIRE) -- Q.ANT, the pioneer in commercial photonic computing, today unveiled the NPU Gen 2+, the latest generation of its Native Processing Unit (NPU) technology, at SC26 in Chicago.
The NPU Gen 2+ adds optical Discrete Fourier Transforms (DFTs), which expands the set of mathematical functions developers can run directly in light with photonic computing. With it, developers can build faster, leaner AI models and high-performance computing (HPC) applications.
Advanced functions like the DFT let AI models work with fewer parameters and less data. On digital hardware, however, computing the transform itself with high data volumes costs time and energy. Because the NPU Gen 2+ delivers DFTs natively in light, Q.ANT demonstrates the potential algorithmic and performance advantages of photonic computing already now.
Developers can now approach computationally intensive functions in new ways using the DFT capabilities of the NPU Gen 2+: Light carries out complex computations in a single pass, extending the set of available instructions beyond add and multiply functions of classical computing and making AI models more expressive and more powerful. For data center operators, integrating photonic hardware into their environments opens new capabilities for users while potentially lowering energy consumption and operating costs significantly.
Today’s language models rely on attention mechanisms that require massive compute to model relationships between tokens, but Q.ANT’s NPU Gen 2+ gives developers a new way to explore these mechanisms leveraging photonic computing as an expansion to classical computation.
At SC26, Q.ANT demonstrates this potential in a hybrid computing system coupling the NPU Gen 2+ with an AMD GPU to run a large language model (LLM) for the first time with optics. The demonstration shows how these mathematical functions performed in light can replace a computationally intensive attention layer and address a major computational bottleneck in generative AI.
Rethinking AI Algorithms for a New Computing Architecture
The NPU Gen 2+ performs these DFTs in a single operation using the physical properties of light, instead of the sequence of energy-intensive electronic calculations that classical computing demands. These advances help establish photonic processors as a complementary computing architecture, with the NPU as a co-processor alongside CPUs and GPUs, so developers can distribute the specific workloads that are most computationally intensive, to where they can be processed the most efficiently.
At the heart of the NPU Gen 2+ is an optical component on the photonic chip that allows the processor to directly perform DFTs, a fundamental mathematical operation used in signal and image processing, simulations, and data analysis. The NPU Gen 2+ extends photonic computing to workloads that are typically expensive for electronic processors. It builds on the processor generation NPU Gen 2, launched in 2025, and its focus on matrix multiplication and non-linear operations.
“With Gen 2+, we’re opening the door to software that uses what light does naturally. That allows leaner AI models with fewer parameters and less data, and that is what will make the next wave of AI affordable,” said Michael Förtsch, CEO of Q.ANT. “With the DFT performed natively, developers now have a new tool to reshape the mathematics of AI models. This is an important step on our roadmap to combine more optical components on our chips for even more sophisticated functions, driving new algorithms and applications. By running a large language model in a hybrid system with an AMD GPU, we have shown that our processors work as co-processors with GPUs. We’re building a community with pioneering developers on the journey to a new era of computing.”
Giving Developers New Tools to Compute with Light
The introduction of the NPU Gen 2+ also expands the potential of Q.ANT’s software development ecosystem.
In September 2026, Q.ANT launched its open-source software development kit (SDK), providing developers with APIs for C and Python, example applications, documentation, and a CPU-based simulation environment. The toolkit allows programmers to begin designing and testing applications for photonic computing without requiring direct access to optical hardware, while the company’s recent launch of its photonic pilot with IONOS allows access to real hardware.
Q.ANT is now extending its SDK to support NPU Gen 2+ capabilities, enabling developers to experiment with Fourier-based algorithms and investigate how these operations can be incorporated into AI and HPC software.
“Photonic computing gives software developers a chance to rethink the mathematics behind their applications,” said Utz Bacher, VP of Software at Q.ANT. “When a complex mathematical operation can be performed directly through the physics of light, it changes the questions developers can ask. Instead of simply optimizing existing code, we can explore different algorithms that use these native operations to solve problems in new ways. Our work with Fourier-based attention layers in language models is an early example of that opportunity.”
Building the Foundation for Hybrid Photonic Computing
The NPU Gen 2+ represents another step in Q.ANT’s roadmap toward more advanced photonic computing architectures.
The new processor retains the integration through PCIe into standard rack-mountable server systems for ease of integration, as the future of computing will depend on the evolution of hybrid systems. Optical processors like Q.ANT’s NPU Gen 2 and Gen 2+ complement the capabilities of CPUs and GPUs with their ability to perform select mathematical functions. Q.ANT plans to offer commercial systems combining Gen 2 and Gen 2+ processors in the second half of 2027.
Going forward, Q.ANT is exploring how multiple optical functions can be combined on its chips to support increasingly complex computational workflows. By expanding the mathematical functions available to developers, it aims to create the foundation for algorithms and applications that optimize data processing across AI and HPC with native photonic computation.
About Q.ANT
Q.ANT is commercializing photonic accelerators for AI and high-performance computing, offering a scalable alternative to transistor-based systems. Its Native Processing Units (NPUs) perform mathematical operations directly on Thin-Film Lithium Niobate (TFLN)-based photonic integrated circuits, enabling energy-efficient, high-performance co-processing for complex computational tasks.
Q.ANT was founded in 2018. Its European headquarters is in Stuttgart, Germany, where it operates its own TFLN chip pilot line, and it maintains an office in Munich and a U.S. office in Austin, Texas. The company secured Europe’s largest Series A for photonic computing in 2025 and introduced its second-generation processor the same year. Its systems are now installed and running at leading scientific computing centers, and commercial deployment.
Media Contacts:
LaunchSquad for Q.ANT
qant@launchsquad.com
Jörg Kochendörfer
joerg.kochendoerfer@qant.gmbh
+49 160 5619730
Photos accompanying this announcement are available at:
https://www.globenewswire.com/NewsRoom/AttachmentNg/69719b55-50da-48ed-9bfd-e823f050510d
https://www.globenewswire.com/NewsRoom/AttachmentNg/6b5100c0-0d13-4025-9d3f-53069c8175a7