Chipta is founded on a simple conviction: the substrate of AI compute is still being invented. We start with photonic acceleration for matrix operations, then study the adjacent architectures that could define the compute layer after this one.
Our work spans waveguide geometry, photonic-electronic control, compiler mapping, and the systems questions required to turn optical matrix operations into useful accelerators — and the same foundation opens a path toward neuromorphic and quantum-inspired compute.
These are research tracks, not product claims. We pursue them in the open, with partners, and only as far as the physics and the market genuinely justify.
Brain-inspired, event-driven architectures are a natural adjacent track for optical systems. Spiking, sparse computation maps well onto analog photonic-electronic substrates — potentially delivering large efficiency gains for always-on and adaptive inference where dense digital arrays are wasteful.
The same optical discipline that makes photonic accelerators possible gives Chipta a credible path to study quantum-inspired methods today and, over a longer horizon, quantum photonic compute. Photons are a natural carrier for quantum information, and the packaging and control we build for accelerators compounds here.
The positioning stays sharp: photonic AI accelerators first. Longer-horizon work moves through neuromorphic photonics and quantum-inspired optical systems.
Photonic AI accelerators for optical matrix multiplication and high-throughput inference.
In developmentEvent-driven, brain-inspired compute built on sparse photonic-electronic signal paths.
ResearchQuantum-inspired and quantum photonic systems for future AI workloads.
ResearchWe collaborate with labs and researchers across photonics, neuromorphic systems, and quantum optics. If our horizons overlap, get in touch.
Talk to us