Q.ANT gives away the software for its light-powered AI chips in a CUDA-style bet on developers

Q.ANT's Light-Powered AI Chips and Software Release

Q.ANT, a startup out of Stuttgart, Germany, builds processors that use light instead of electricity to perform some of the math behind AI. The company positions its chips as a way to run AI on a fraction of the power today's chips require. Unlike traditional approaches, Q.ANT pushes a free, open-source software kit to GitHub this week that enables developers to build and test programs on a normal computer before running them on the actual chips once access is available.

The Q.ANT Native Computing Toolkit

The released toolkit, called the Q.ANT Native Computing Toolkit, is free on GitHub under a license permitting commercial use. Developers can work in Python or C. A key feature is a simulator that mimics the chip on a regular computer, requiring no Q.ANT drivers. This approach mirrors Nvidia's strategy but focuses on photonic computing rather than GPU software ecosystems.

Current Capabilities

The first version of the toolkit focuses on running pre-trained AI models. Examples include reading handwritten numbers, identifying objects in photos, and outlining shapes in images. However, training still occurs on conventional CPUs and GPUs. Notably, there is currently no plugin for PyTorch-the most popular AI framework-so developers must wire components together manually, similar to the example implementations provided.

The Power Advantage

The primary selling point for photonic computing lies in power efficiency. Traditional AI chips consume significant energy moving data between memory and the processor. Q.ANT's chips handle portions of the computation using light-specifically wave-shaped functions resembling a cosine-which regular chips calculate digitally. According to Q.ANT, AI models built around these light-based functions achieve better results with fewer parameters, resulting in smaller models, reduced data movement, and lower power consumption.

Founder Perspective

Michael Förtsch, Q.ANT's founder and CEO, describes the release as the "Linux moment" of photonic computing. He emphasizes that "an ecosystem isn't created by hardware alone. It emerges when the software layer is open and others can build on it."

Company Background

Q.ANT recently raised €62 million in July 2025 in a funding round led by Cherry Ventures, UVC Partners, and imec. Earlier in 2026, the company reported that its second-generation chips were operational at the Leibniz Supercomputing Centre near Munich. Independent tests at that facility showed the new chip was more than 50 times faster at the type of mathematical operations central to AI models and used six times less energy on typical tasks compared to previous generations. Larger claims-such as up to 30 times better energy efficiency-have not specified what they are measured against.

Competitive Context

Good software alone cannot sustain a new chip's success. Nvidia has maintained dominance with CUDA, a developer framework for its GPUs, for nearly two decades. Similarly, Graphcore, a British AI chip startup, developed its own software kit before being acquired by SoftBank in 2024. Q.ANT positions its open-source approach as distinct from these paths. While other players like Xanadu have offered free, open software for light-based quantum computers since 2018, Q.ANT highlights its toolkit as the first openly available software kit specifically designed for programming a photonic processor.