15–20 Nov 2026
McCormick Place Convention Center
Meet Q.ANT at SC26.
Experience analog photonic processing, demonstrated on site with running AI models, promising energy and cost efficiency for AI and HPC.
AI with the speed of light.
Q.ANT pioneers photonic computing to make AI faster, more energy-efficient and cheaper.
Today’s AI models mostly rely on multiply and add. Light can do more. Our processors compute the complex, non-linear math of AI natively in light, making AI models more expressive and more powerful. That gives developers a new tool for algorithmic innovation: leaner models that need less data and fewer parameters, freeing up compute capacity and addressing energy constraints.
Join us on the journey to the new compute.
More than 50,000 hours
of photonic compute in live data-center operation
4 sites
LRZ, Jülich Supercomputing Centre, KIT and IONOS
> 100
Q.ANTies in Stuttgart, Austin and Munich. Founded in 2018 by Michael Förtsch.
Our Partners
AI’s next leap will not come from more of the same.
The problem
AI is on track to outgrow the world’s power supply.
If today’s growth continues, AI alone would need as much electricity as the entire planet produces within only few years. Faster chips won’t fix that, because most of the energy is not spent on computing but on the huge amount of data moved within the system.
The energy bill
Arithmetic (MAC)~1 pJ
On-chip SRAM read~1–3 pJ
Fetch from DRAM>100 pJ
Move across the rack~500 pJ
Why we exist
The problem
If moving data is the real cost, the answer is to move less of it. Today’s AI algorithms break the rich math of AI into countless multiplications and additions to accommodate common technology limitations.
The solution
Photonic computing performs non-linear functions natively, in a single pass. Denser math shrinks the data footprint, and it opens a design space rarely explored: leaner models, new architectures, new ways to learn. Less data movement means less memory and less energy required.
Where you can make a difference.
No one has written the playbook for photonic AI yet. We are writing it in the open, with researchers, partners and developers who want to build what comes after.
Time-series prediction
Non-linear forecasting models run on the NPU with 65% fewer parameters at the same prediction accuracy and with significantly less compute.
- Sales and demand forecasting in retail
- Predictive maintenance in industry
- Load management and grid frequency monitoring
Industrial vision
Defect detection and classification models achieve the same detection performance on the NPU with 40% fewer parameters, for faster inline inspection.
- Quality assurance and defect detection
- OCR and document classification
- Real-time object detection
LLM and foundation model inference
The NPU accelerates linear and non-linear operations of large language models in a single optical step, without the bottlenecks of conventional GPU inference pipelines.
- Inference with open-source LLMs
- Embeddings and semantic search
- High-throughput RAG pipelines
Physical AI and robotics
Control and sensing models benefit from the low latency of photonic processing, wherever response time matters.
- Robot control and motion planning
- Sensor fusion and environment modelling
- Autonomous systems with real-time requirements
Scientific simulation
The NPU natively accelerates Fourier transforms and partial differential equations, without the detour via conventional semiconductor architectures.
- Climate and fluid-flow simulation
- Quantum physics and materials research
- Industrial R&D with heavy simulation workloads
World models
World models need non-linear operations at scale. That is exactly what the photonic architecture of the NPU Gen 2 is designed for.
- Simulation-based reinforcement learning
- Generative environment models
- Autonomous agents with a planning horizon
Get the SDK. Code for light.
New math. New models. Your move.
Only multiply and add was yesterday’s toolbox. The Q.ANT SDK gives you native non-linear functions, computed with light. Design architectures around richer operations instead of approximating them, build models that do more with fewer parameters and less data. This is where algorithmic innovation begins: new model architectures, new network designs, new ways to rethink computing.
Get the SDK. Run your first models, built for light. Show us what’s possible.
The SDK includes
- APIC and Python
- Sample applicationsImage learning, image generation, time series prediction, object detection
- DocumentationFrom native functions to high-level operations
- Simulation backendRuns on a standard CPU, so developers can build and validate applications on their own systems
Get access to our whitepapers Early AI applications on Q.ANT’s photonic NPS
Test photonic computing on photonic hardware
IONOS × Q.ANT launch the world’s first Photonic Computing Pilot.
First access for companies, researchers, and developers to Q.ANT’s photonic processor, to explore Photonic Computing.
Apply for photonic hardware access
International applicantsPilot program in EnglishBewerbung aus DeutschlandZur deutschen IONOS-SeiteGet your hands on deep tech.
We think outside the CMOS box. So we built a new one: a photonic server right next to the servers you already know.
At the core of our servers, our Native Processing Unit (NPU) runs on light, giving you access to complex, non-linear functions natively implemented on the chip.
- Available now
NPU Gen 2
Second generation, with 8 optical cores. Enhanced analog units for nonlinear functions.
- Coming soonHW v2.1
NPU next Gen
More optical elements for more efficiency.
Available as proof of concept. Comes as Native Processing Server NPS in 19-inch server format, plug and play, no cooling required.
Insights from the frontier of photonic computing.
Research, collaborations, developer notes and announcements, written by the people building it.
Latest news
Press releaseQ.ANT Launches World First Open-Source SDK for Photonic ComputingQ.ANT, the pioneer in commercial photonic computing, today debuted its open-source software development kit for photonic computing, making the programming tools for a photonic processor publicly available for the first time.Read more 
Press releaseQ.ANT Appoints IBM Veteran Joerg Behrend to Lead Photonic Processor Hardware DevelopmentQ.ANT, a leading provider of photonic processors for energy-efficient high-performance computing and artificial intelligence, is strengthening its leadership team with the addition of Joerg Behrend, who joined the company as Vice President …Read more 
Press releaseQ.ANT Runs Generative AI on Photonic HardwareQ.ANT, the pioneer in commercial photonic computing, today demonstrated the first complex, production-relevant AI workloads on its photonic hardware.Read more Upcoming event
Latest from Q.ANT on LinkedIn
Q.ANTQ.ANT has won The Spark 2026, the technology award by Handelsblatt and McKinsey & Company 🏆This year's theme was “Scaling Deep Tech” - and that is what our team works on every day: turning photonic computing into scalable computing infrastructure for more…Read on LinkedIn 
Q.ANTLet’s talk breakthroughs at SC26Meet Q.ANT at #SC26. Explore how to put light to work in photonic computing to improve efficiency in HPC and AI. Power, not compute, now limits the growth of AI. Start…Read on LinkedIn 
Q.ANTOur photonic computing hardware is already running in leading German supercomputing centers.Meet our team at #SC26 across three partner booths at HLRS - High-Performance Computing Center Stuttgart (booth #4106), Leibniz Supercomputing Centre (booth #4254) and…Read on LinkedIn Latest from Dr. Michael Förtsch on LinkedIn
Dr. Michael FörtschIn 2018, Q.ANT GmbH was a handwritten two-page business plan.Last night in Munich, it won The Spark 2026. This year's theme: Scaling Deep Tech. I pitched our product vision in two lines: → AI will be stopped by power costs, not…Read on LinkedIn 
Dr. Michael FörtschI went to Jena to talk about the physics of photonic computing.The best conversations afterwards were about packaging. Last week I opened the JENOPTIK Innovation Days at the invitation of Dominic Dorfner and Adrian König - 150…Read on LinkedIn 
Dr. Michael FörtschThree market reports. Same category. Same year, 2035.$22 billion. $33 billion. $41 billion. The $41 billion one is in my feed right now. It names Q.ANT among the key players, so I have every reason to like that number 😀…Read on LinkedIn Join the team.
Series A funded. Chip deployed. Grow team.
Our photonic processors already run in leading data centers, built end to end from our own Stuttgart pilot line to the SDK. With our Series A in place, we’re looking for the people who take photonic computing to scale, in Stuttgart, Munich and Austin.
- StuttgartHeadquarters and pilot line
- MunichOffice
- AustinOffice
Let’s talk!
Questions about developing applications, the NPU, computing time, the SDK, a partnership or how to back Q.ANT? Write to us.