AI unit
Built on Arago’s proprietary multi-physics computing core, high-volume AI tasks with 𝑛^3 complexity are accelerated and executed at ultra-low power consumption.
Arago is leading the shift to optical computing, to drive the course of history forward.
Scroll to discover
AI is boxed in by the limits of today's hardware
Modern AI models are increasingly constrained by the compute and memory ceilings of traditional silicon-based accelerators. Advancing AI efficiency and deployment implies moving beyond traditional transistor-based technology.
Arago has built a photonic AI accelerator optimized for energy efficiency and high throughput. By leveraging light as a computing principle, the chip maximizes performance per watt and per dollar, to power AI workloads with significantly lower energy and cost overhead.
Architected to practically deliver 10×–30× reductions in energy consumption
Built on Arago’s proprietary multi-physics computing core, high-volume AI tasks with 𝑛^3 complexity are accelerated and executed at ultra-low power consumption.
Its deterministic architecture maximizes memory transfer efficiency and speeds up pointwise operations, achieving optimal performance for tasks with 𝑛^2 complexity.
Higher computation capacity: a byte of data in a GPU is only used for a few operations, whereas on Arago's system, that same byte enables thousands of operations, achieving multi-PetaOp/s performance.
Higher throughput: excel when treating tensor sizes in the hundreds and thousands.
Lower energy consumption: reduce costs and enable the next multi-die 3D systems.
Compatible with industry-standard software frameworks and hardware platforms
The computing infrastructure should not have to deal with the specificities of a multi-physics-based processor - which is why it’s all abstracted at the instruction level, hidden from the end user.
Arago’s software stack, CARLOTA®, seamlessly interfaces PyTorch and abstracts away low-level hardware details, enabling developers to deploy and scale AI models without modifying their existing codebase.
Driving the course of history forward
If you're excited about shaping the future of computing and AI, we'd love to hear from you
The power of computing has enabled humanity to explore new planets, discover treatments for the rarest diseases, and produce the most beautiful films ever seen.
We don’t want discoveries and creativity to be slowed by a lack of appropriate infrastructure.
High velocity
Speed matters. We move quickly, one step at a time.
One unit
We’re all in this together, with relationships grounded in trust, respect, and camaraderie.
Do great things
We deliver work we’re proud to sign our name to.
We are a team of AI engineers and physicists who believe in great science and fast achievements. We’re looking for bold jack-of-all-trades who love to ship products quickly. If that’s you, please apply for one of our open roles here.
Position | Location | |
---|---|---|
Senior Hardware Design Engineer | Mountain View, Paris | |
Mixed-Signal IC Design Engineer | Tel Aviv, Paris | |
Senior Optical Co-Integration Engineer | Paris |
PRACTICAL OPTICAL COMPUTING
00%
Loading