Electronics Weekly
“For the first time, attendees will be able to interact directly with functional photonic computing, witnessing how light can drive energy and computational efficiency for AI, physics simulations, and other complex scientific workloads”. according to the company. “Q.Ant is attacking two of the biggest challenges in photonic computing: integration and precision. The Native Photonic Server promises to deliver up to 30 times the energy efficiency of conventional technologies.”
It claims:
- 16-bit floating point precision with 99.7% accuracy for all computational operations on the chip
- 40–50% fewer operations required for equivalent output
- No need for active cooling infrastructure
At the system’s core is a proprietary thin-film lithium niobate photonic chip designed to execute non-linear mathematical operations directly in the optical domain.
“Its analogue processors excel at non-linear and mathematical operations that digital systems struggle with”, said Q.Ant. “By computing non-linear functions and Fourier transforms directly with light, the server reduces the number of parameters required in AI models, simplifying architectures and lowering overall system demand.”
Connections are through standard PCI Express and industry-standard AI frameworks including TensorFlow, PyTorch and Keras arew supported.
Find Q.Ant’s demonstration on stand G12 in hall H at ISC 2025 – 10 – 12 June in Hamburg, Germany.
