Founder Name
Placeholder — add background: prior company, domain, what they're driving at Chipta.
Chipta is a research-driven startup developing photonic AI processors. We believe the substrate of AI compute is still being invented — and that light is the most promising path through the data-movement walls now limiting large models.
Inference demand is outgrowing what electrical data movement can economically deliver, just as silicon photonics reaches manufacturing maturity. That intersection is a narrow window to build something genuinely new.
We start where photonics is strongest: optical matrix multiplication for high-throughput AI inference. From that foundation, we study the adjacent architectures — neuromorphic and quantum-inspired — that could define the compute layer after this one.
We are early, deliberately focused, and building in close collaboration with the AI labs, semiconductor partners, and research groups who will use and shape this hardware.
We win one defensible problem — optical matrix multiplication for inference — before we widen. Focus is the advantage.
We let the device physics lead the architecture, not the other way around. The roadmap follows what light is genuinely good at.
Real accelerators are co-designed with the labs and fabs that will use and make them. We validate in the open, with partners.
Placeholder profiles — replace with real bios.
Placeholder — add background: prior company, domain, what they're driving at Chipta.
Placeholder — add background: photonics / chip architecture experience, key prior work.
Placeholder — add background: lab, publications, area of expertise in silicon photonics.
Placeholder — add background: systems / compiler / hardware experience.
If you want to build computing hardware from first principles — across photonics, mixed-signal, compilers, and systems — we'd like to hear from you.
Get in touch