Vendors & Products
30 entriesProducts and vendors listed for AI Chip Providers.
01Google Cloud TPUGoogle's custom-built AI accelerator ASICs (Tensor Processing Units), spanning cloud-scale training to edge inference. Google describes Ironwood as its seventh-generation TPU, engineered for inference.
02Arm AIArm's AI-optimized CPU, GPU, and NPU IP powers the vast majority of the world's mobile and edge AI silicon.
03Apple SiliconApple's A-series and M-series chips ship a dedicated Neural Engine for on-device machine learning across iPhone, iPad, and Mac.
04Intel AIIntel's AI portfolio spans Xeon CPUs, Gaudi accelerators, and NPUs to help enterprises build and deploy AI at scale.
05AMD AI SolutionsAMD's Instinct GPUs, Ryzen AI NPUs, and (following the Xilinx acquisition) adaptive FPGA compute for AI workloads.
06GraphcoreGraphcore's Intelligence Processing Units (IPUs) are purpose-built processors designed specifically for machine intelligence. SoftBank Group completed its acquisition of Graphcore in July 2024, and Graphcore continues as a subsidiary.
07MediaTek AIMediaTek builds AI-integrated system-on-chip products powering smartphones, smart homes, and connected devices.
08NVIDIA AINVIDIA's GPUs remain the dominant compute platform for AI training and inference, from the datacenter to the edge.
09MythicMythic builds analog compute-in-memory processors (AMPs) that deliver efficient AI inference at the edge. Mythic restructured in 2022 and announced a US$125 million funding round in December 2025; its efficiency claims are its own.
10Qualcomm AI ResearchQualcomm's Hexagon NPUs and AI research push efficient, on-device intelligence across billions of Snapdragon-powered devices. Qualcomm has also announced AI200 and AI250 inference accelerators for data centres, which it says will be available in 2026 and 2027, with Humain as its first named customer.
11BlaizeBlaize designs low-power AI processing chips purpose-built for edge inference in vision and video applications. Blaize became a publicly listed company through a SPAC merger in January 2025, according to its SEC filings.
12TSMCTSMC is the leading semiconductor foundry manufacturing AI chips for NVIDIA, Apple, AMD and others, using AI-driven smart manufacturing itself.
13Samsung ExynosSamsung's Exynos system-on-chip line integrates dedicated NPUs for on-device AI in mobile and consumer electronics.
14HiSiliconHiSilicon (Huawei's chip design arm) develops the Kirin and Ascend series of AI-capable processors.
15Texas InstrumentsTexas Instruments supplies embedded processors and microcontrollers with AI acceleration for industrial and edge devices.
16VIA TechnologiesVIA Technologies builds AI-enabled SoC solutions for automotive, edge, industrial, and building automation applications.
17Imagination TechnologiesImagination's PowerVR Vision & AI cores deliver neural network acceleration for mobile, automotive, and smart-camera markets.
18Huawei AscendHuawei's Ascend series (including the Ascend 910) are high-performance AI processors for training and inference at scale.
19Microsoft Project BrainwaveMicrosoft Research's FPGA-based deep learning platform for real-time AI inference in the cloud and at the edge. Announced as a Microsoft Research project in 2017, so check Microsoft's current Azure documentation for availability.
20GroqGroq's LPU (Language Processing Unit) is custom silicon for fast AI inference. Groq's site now presents the company as an inference cloud, and press reports say Nvidia took a non-exclusive licence to Groq's technology and hired its leadership in December 2025.
21Kalray AI & ComputeKalray's MPPA massively parallel processor arrays capture and analyze high-throughput data flows in real time.
22AWS InferentiaAWS's custom-designed Inferentia accelerators deliver high-performance, low-cost machine learning inference on Amazon EC2.
23CerebrasWafer-scale AI processors and an inference and training cloud built on them. Cerebras began trading on Nasdaq as CBRS on 14 May 2026, according to press reports.
24SambaNovaAI inference platform built on its reconfigurable dataflow chips. Press reports in late 2025 described Intel acquisition talks that did not complete; Intel instead took part in a US$350 million funding round in February 2026.
25TenstorrentRISC-V and AI accelerator hardware and software.
26AWS TrainiumAWS's custom accelerators for deep learning training and inference.
27Microsoft Azure MaiaMicrosoft's custom AI accelerator for Azure, described in its announcement blog.
28Meta MTIAMeta's in-house Meta Training and Inference Accelerator family, which Meta says it is developing in several generations for ranking, recommendation and generative AI workloads.
29PositronAI inference hardware startup building accelerators for running large language models in data centres.
30SiMa.aiEdge AI company whose Modalix and MLSoC chips combine ML processing with application processors for physical AI at the edge.
More AI Accelerator Vendors & Benchmarks
8 entriesFurther vendors and benchmark bodies for AI silicon.
31NVIDIA Data Center GPUsGPUs for training and inference.
32AMD InstinctAMD data center accelerators.
33EtchedTransformer-specific accelerator startup.
34d-MatrixIn-memory compute for inference.
35LightmatterPhotonic computing for AI.
36RebellionsKorean AI chip company.
37Axelera AIEuropean edge AI accelerators.
38MLPerf BenchmarksIndustry benchmarks for AI hardware.
Benchmarks & Platform Docs
2 entriesHardware benchmarks and platform docs.
Related Pages
5 entriesRelated directories on this site.
41GPU Cloud ProvidersRelated directory on this site.
42Edge AIRelated directory on this site.
43AI Chip CompilersCompilers and toolchains for AI accelerators.
44Physical AIRelated directory on this site.
45CPU, GPU, TPU, NPU, LPU, DPU & QPU ExplainedWhat each processor type is built for, how they work together in an AI stack, and how to choose between them.