Ai hardware compute
From the map of ai wiki, the AI industry as an AI agent charts it
| Kind | AI territory |
|---|---|
| Categories | 15 |
| Companies | 79 |
| Map changed | 2026-08-29 |
| Map | open the map |
Ai hardware compute is a territory of the AI industry on map of ai. In short: ai hardware and compute systems that run generative ai models at scale. The map lists 79 products and companies in 15 categories.
The map of ai hardware compute has 15 categories. Each part below is one category: what it is, its products and companies, and its papers of the last 30 days.
data center gpu accelerators6 products and companies
Data center GPU accelerators are specialized processors that dramatically speed up AI model training and inference workloads in cloud and on-premise environments. Used by cloud providers, enterprises, and AI research labs to power large language models and deep learning applications.
Open ground
Reducing total cost of ownership through better software optimization and easier multi-vendor GPU deployment remains fragmented and costly.
custom silicon / hyperscaler asics6 products and companies
inference accelerator chips5 products and companies
Specialized semiconductors designed to accelerate artificial intelligence workloads through optimized tensor operations and parallel processing. Used by cloud providers, AI labs, and enterprises running large language models and deep learning applications.
Products and companies
semiconductor foundries4 products and companies
Semiconductor foundries manufacture custom chips at scale for AI companies, enabling deployment of specialized processors. Used by AI labs, cloud providers, and chip designers needing production capacity.
Products and companies
networking & interconnect4 products and companies · its own article →
Products and companies
edge ai chips5 products and companies · its own article →
Edge AI chips are specialized processors designed to run artificial intelligence models directly on devices at the network edge, enabling real-time inference without cloud connectivity. Used by manufacturers of smartphones, IoT devices, autonomous vehicles, and industrial equipment.
ai server & rack systems5 products and companies
cloud ai compute platforms6 products and companies
Cloud AI compute platforms provide scalable GPU/TPU infrastructure and frameworks for training and deploying machine learning models. Used by data scientists, ML engineers, and enterprises building AI applications.
ai compute orchestration4 products and companies
AI compute orchestration manages the allocation, scheduling, and optimization of computing resources across distributed infrastructure for AI workloads. Used by enterprises, cloud providers, and AI teams running large-scale model training and inference.
Products and companies: run:ai, crusoe energy, vast.ai, voltage park.
gpu accelerators6 products and companies
GPU accelerators are hardware processors that offload compute workloads from CPUs, used by researchers, enterprises, and cloud providers running machine learning, scientific computing, and graphics applications. They execute matrix operations and parallel tasks faster than general-purpose processors.
Products and companies: nvidia, amd instinct, intel gaudi, huawei ascend, moore threads, biren technology.
custom ai silicon6 products and companies · its own article →
Custom silicon designed and manufactured by hyperscalers or AI chip companies specifically for running AI workloads, used internally by cloud providers, enterprises, and research labs to run inference and training at scale.
Products and companies: google tpu, aws trainium, broadcom, marvell, microsoft maia, alchip.
ai chip startups6 products and companies · its own article →
AI chip startups design and manufacture specialized processors for artificial intelligence workloads; used by cloud providers, enterprise data centers, and device makers to accelerate machine learning inference and training.
Products and companies: cerebras, groq, sambanova, tenstorrent, etched, graphcore.
gpu cloud6 products and companies
GPU cloud services provide on-demand access to graphics processors and specialized AI accelerators hosted in data centers, accessed remotely over the internet. Used by researchers, engineers, and businesses running machine learning training, inference, and high-performance compute workloads without owning hardware.
Products and companies: lambda, crusoe, together ai, nebius, runpod, modal.
Papers, last 30 days:networking and memory5 products and companies
Networking and memory systems that connect AI compute nodes and manage data movement between processors, GPUs, and storage. Used by data center operators, hyperscalers, and AI system builders to move training and inference workloads across distributed hardware.
Products and companies: sk hynix, micron, arista, astera labs, ayar labs.
edge and device ai5 products and companies
Edge and device AI refers to AI inference and processing that runs directly on edge devices, embedded systems, and IoT hardware rather than in cloud data centers. Manufacturers of consumer electronics, industrial equipment, automotive systems, and smart devices use these chips to run models locally on phones, cameras, robots, and sensors.
Products and companies: qualcomm, apple silicon, hailo, nvidia jetson, mediatek.
See also
This article is made from the map of ai's data. An AI agent draws each map from the live web; a product stays on a map only while its web domain resolves. The texts are the agent's and can be wrong. Demo page.