Chips

High-performance chips and open software, advancing AI from data centers to the edge.

Frontier inference clusters, co-designed from chips to software.

Computing from chip design to manufacturing.

Positron makes purpose-built hardware to accelerate generative AI.

Efficient AI inference hardware and software for edge devices and servers.

Inference hardware and software designed for practical data-center power limits.

Rethinking the limits of AI, Lightmatter merges photonics and computing to build a future where speed, efficiency, and intelligence converge

Compute-in-memory technology and IP for energy-efficient local AI.

Wafer-scale computers and cloud services that accelerate AI training and inference.

AI processors and software designed together, pairing massively parallel IPUs with Poplar for machine learning.

Data infrastructure technology for enterprise, cloud, automotive, and carrier architectures.

Purpose-built inference silicon and systems for enterprise and sovereign AI.

Memory-centric compute for AI inference.

Groq is the premier neocloud for fast inference.

Analog compute-in-memory hardware and software for efficient AI inference.

A chip-to-model platform for fast, efficient AI inference.

Analog in-memory AI acceleration for local devices and workstations.

AI inference and vision processors for efficient local intelligence.

Chips, systems and software co-designed to power AI factories and the physical world.

Hardware and software for complete physical AI applications at the edge.