Comprehensive Overview of Global AI Chip Companies and Their Representative Products
An extensive survey of the global AI chip market lists over 90 companies—including major manufacturers, Chinese startups, and traditional IP vendors—detailing each firm's representative AI processors, development history, and key technological features to illustrate current industry trends and competitive landscape.
The AI chip market is intensifying, with ARK research indicating that more than 60 companies worldwide are in advanced stages of developing or selling specialized processors to accelerate AI applications, spanning the United States, China, Europe, and Israel.
China leads the AI chip development wave, with many non‑traditional chip players such as Google and Amazon entering AI chip design, fostering a trend toward custom silicon and a resurgence of optical computing.
Chip expert Dr. Tang Shan maintains a GitHub channel called "AI Chip Panorama" that visualizes a list of 99 AI‑related companies, providing a convenient reference for the global AI chip ecosystem.
Major IC manufacturers and their representative AI chips include Intel (Nervana, Mobileye EyeQ, Movidius, Arria 10 FPGA, Loihi), Samsung (Exynos 9825 with integrated NPU), NVIDIA (Volta, Turing, T4, Xavier, NVDLA), Qualcomm (Snapdragon 855, Cloud AI 100), AMD (EPYC), Xilinx (Versal), MediaTek (Dimensity 1000, Helio P90), and others, each with brief histories and product highlights.
Technology giants and HPC vendors also contribute AI silicon: Google (TPU, Edge TPU), Microsoft (Brainwave FPGA‑based accelerator), Amazon (Inferentia), Apple (A11/A12/A13 Neural Engine), IBM (TrueNorth, Watson), Huawei (Ascend 910/310, Kirin 980/990 5G, Hi3559A), Baidu (Kunlun, Honghu), Alibaba (Hanguang 800, TG6100N), and Tesla (Full‑Self‑Driving chip).
Chinese AI‑chip startups are numerous, including Cambricon (MLU series), Horizon Robotics (Journey, Sunrise), Bitmain (BM1680/1682/1684/1880), CloudWalk (DeepEye 1000), Yitu (QuestCore), Zhaoguan Electronics (N171), Canaan (K210), CoolChip (AR9000), Tanshen (Voitist 611), Qingwei (TX101/210/510), Yizhi (TAi8010), Black Sesame (Huashan One), Suiyuan (DTU SuiSi), Zhicun (MemCore001), Qitai (CI100X/110X), Speech‑AI (TAIHANG), Yunzhisheng (UniOne), Mobvoi (A1), Rokid (KAMINO18), among others.
International startups expanding the AI‑chip landscape include Cerebras Systems (CS‑1), Graphcore (GC2), Habana Labs (Gaudi, Goya), Hailo (Hailo‑8), Blaize (XPlorer X1000), Kalray (MPPA series), PEZY Computing (SC2), Eta Compute (ECM3531), Greenwaves (GAP8), Gyrfalcon Technology (Lightspeeur, GAINBOARD), Preferred Networks (MN‑Core), InnoGrit (Shasta, Rainier, Tacoma), and Kneron (KL520).
Traditional IP vendors are also adapting to AI demands: Arm (DynamIQ, Compute Library), Synopsys (DesignWare EV6x vision processor), Imagination (PowerVR Series2NX NN accelerators), CEVA (NeuPro NN accelerator), Cadence (Tensilica Vision DSP), Chipone (VIP8000), and videantis (v‑MP6000UDX), providing IP, tools, and design services for AI chips.
The article concludes with a disclaimer that the content is reproduced with attribution and invites readers to access additional technical resources via the original source.
Architects' Tech Alliance
Sharing project experiences, insights into cutting-edge architectures, focusing on cloud computing, microservices, big data, hyper-convergence, storage, data protection, artificial intelligence, industry practices and solutions.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.