Decart

Decart develops real-time generative AI infrastructure and world models for live video, interactive experiences, physical AI, and AI inference.

Decart is an AI research and infrastructure company developing real-time generative AI models and the computing stack required to run them with low latency. Its technology includes DOS, an optimized AI inference and training stack, along with Lucy and Oasis world models. These systems support live video transformation, interactive digital environments, immersive experiences, and simulations designed for physical AI and robotics applications.

MemryX

Learn about MemryX, a semiconductor company developing edge AI accelerators for efficient neural network inference across devices and edge systems.

MemryX is a fabless semiconductor company that develops AI accelerators for running neural network inference at the edge. Its MX3 accelerator uses a dataflow architecture designed to offload AI processing from host processors and efficiently handle streaming inputs from cameras and other sensors. MemryX technology supports scalable edge AI deployments across embedded systems, industrial equipment, transportation, robotics, intelligent infrastructure, and edge servers.

Quadric

Learn about Quadric, a semiconductor IP company developing programmable GPNPU processor technology for edge AI inference and AI-enabled chips.

Quadric is a semiconductor IP company that develops programmable neural processing technology for AI inference in chips and edge devices. Its Chimera GPNPU architecture combines neural processing and general-purpose programmability, allowing semiconductor and systems companies to integrate AI processing into custom SoCs. The technology supports workloads including computer vision, ADAS, industrial vision, robotics, smart cameras, consumer devices, and other on-device AI applications.

Blaize

Learn about Blaize, an AI computing company developing edge AI processors, accelerator platforms, and software for real-time AI inference applications.

Blaize is an AI computing company that develops processor hardware and software for deploying AI inference from edge devices to data center environments. Its Graph Streaming Processor architecture is designed for energy-efficient, low-latency AI and machine learning workloads, while its hardware platforms and AI software suite support application development and deployment. Blaize technology is used across industrial, smart infrastructure, retail, automotive, enterprise, and other real-time edge AI applications.

Groq

Groq is an AI company that designs hardware and software for AI inference, which is the process of running generative AI models.

Groq is a technology company that builds specialized AI hardware and software designed to speed up the performance of large language models. Its custom-built chips and compiler stack are engineered for ultra-low latency, enabling real-time processing for demanding AI tasks such as natural language generation and machine learning inference. Groq’s solutions are used in applications where speed and efficiency are critical.