Chips & Compute

AI accelerators, inference silicon and the compute platforms everything else runs on.

18 tools · 58 changes / 30d

AMD Instinct MI355X

AMD

High-memory GPU alternative

Version: MI355X
Cost: Enterprise hardware, pricing not published
Model: CDNA

Apple Silicon Neural Engine

Apple

On-device inference across Mac, iPhone and iPad.

Version: M5
Cost: Bundled with hardware
Model: Apple Foundation Models

AWS Trainium2

AWS

Cloud-native training silicon

Version: Trainium2
Cost: Pricing is not published on the page. Available via EC2 instances/UltraServers.
Model: Neuron compiler

Broadcom custom accelerators

Broadcom

Co-designed XPUs behind most hyperscaler silicon.

Version: 2025
Cost: Enterprise supply agreements
Model: powers Google TPU

Cerebras WSE-3

Cerebras

Wafer-scale inference and training

Version: CS-4
Cost: Free trial available with developer tier pricing (includes a one-time $5 promotional credit).
Model: Codex-Spark

d-Matrix Corsair

d-Matrix

Digital in-memory compute for inference efficiency.

Version: Corsair
Cost: System sales
Model: n/a

Etched Sohu

Etched

Transformer-only ASIC betting on architecture lock-in.

Version: Sohu
Cost: System sales
Model: transformer-only

FuriosaAI RNGD

FuriosaAIAsia

Korean inference silicon with strong perf-per-watt.

Version: RNGD
Cost: System sales
Model: n/a

Google TPU Ironwood

Google

Inference-first TPU pods

Version: 7th generation (Ironwood)
Cost: from $5.4/unit
Model: Gemini

Graphcore IPU

SoftBankEU

IPU accelerators now inside SoftBank.

Version: Bow IPU
Cost: System sales
Model: n/a

Groq LPU

Groq

Deterministic low-latency inference

Version: GroqCloud API
Cost: Pricing not published on current page. Custom/usage rates apply.
Model: canopylabs/orpheus-arabic-saudi

Huawei Ascend 910C

HuaweiChina

China domestic AI accelerator

Version: Ascend 910C
Cost: Pricing is not publicly available; enterprise/quote-based pricing.
Model: Ascend 910C

Intel Gaudi 3

Intel

Cost-focused training and inference accelerator.

Version: 1.24.2
Cost: Pricing not publicly listed (contact Intel sales or cloud partners)
Model: Accelerator

Lightmatter

Photonic interconnect for scaling accelerator fabrics.

Version: Passage M1000
Cost: System sales
Model: n/a

NVIDIA Blackwell

NVIDIA

GB200/B200 rack-scale AI systems

Version: GB200 NVL72
Cost: Pricing is not published (enterprise/quote-only)
Model: Blackwell

Qualcomm AI200

Qualcomm

Rack-scale inference accelerators entering the datacentre.

Version: AI200
Cost: System sales
Model: n/a

SambaNova SN40L

SambaNova

Dataflow chips for model serving

Version: SN40L
Cost: Free tier available, pay-as-you-go developer tier, and subscription-based Enterprise plan.
Model: DeepSeek-V3.1

Tenstorrent Blackhole

Tenstorrent

Open RISC-V AI hardware

Version: Blackhole
Cost: from $999/unit
Model: Blackhole Tensix

Recent changes in Chips & Compute

Nov 7, 2026first seenmodel
Groq LPU — Groq LPU added qwen/qwen3-32b to its model stack
LPU v2, Llama / Kimi / GPT-OSS, Llama 3.1 8B Instant, Llama 3.3 70B Versatile, Orpheus English, Qwen 3.6 27B, Whisper V3 Large, canopylabs/orpheus-arabic-saudi, canopylabs/orpheus-v1-english, kimi-k2-instruct-0905, llama-4-maverick-17b-128e-instruct, llama-4-scout-17b-16e-instruct, meta-llama/llama-4-maverick-17b-128e-instruct, meta-llama/llama-4-scout-17b-16e-instruct, minimax-m2.5, minimax/minimax-m2.5, minimaxai/minimax-m2.5, moonshotai/kimi-k2-instruct-0905, openai/gpt-oss-120b, openai/gpt-oss-20b, openai/gpt-oss-safeguard-20b, orpheus-arabic-saudi, orpheus-v1-english, qwen/qwen3-vl-32b-instruct, qwen3-32b, qwen3-vl-32b-instruct, whisper-large-v3LPU v2, Llama / Kimi / GPT-OSS, Llama 3.1 8B Instant, Llama 3.3 70B Versatile, Orpheus English, Qwen 3.6 27B, Whisper V3 Large, canopylabs/orpheus-arabic-saudi, canopylabs/orpheus-v1-english, kimi-k2-instruct-0905, llama-4-maverick-17b-128e-instruct, llama-4-scout-17b-16e-instruct, meta-llama/llama-4-maverick-17b-128e-instruct, meta-llama/llama-4-scout-17b-16e-instruct, minimax-m2.5, minimax/minimax-m2.5, minimaxai/minimax-m2.5, moonshotai/kimi-k2-instruct-0905, openai/gpt-oss-120b, openai/gpt-oss-20b, openai/gpt-oss-safeguard-20b, orpheus-arabic-saudi, orpheus-v1-english, qwen/qwen3-32b, qwen/qwen3-vl-32b-instruct, qwen3-32b, qwen3-vl-32b-instruct, whisper-large-v3
source
Nov 7, 2026first seencapability
AWS Trainium2 — New capabilities: Stochastic Rounding
105 tracked106 tracked · +Stochastic Rounding
source
Nov 7, 2026first seenversion
Google TPU Ironwood — Google TPU Ironwood moved to 7th generation (Ironwood)
Ironwood (7th Generation)7th generation (Ironwood)
source
Nov 7, 2026first seenversion
Cerebras WSE-3 — Cerebras WSE-3 moved to CS-4
WSE-3CS-4
source
Nov 7, 2026first seenmodel
Cerebras WSE-3 — Cerebras WSE-3 added Kimi K3.1 to its model stack
Codex-Spark, GLM-4.7, GPT 5.6 Sol, GPT-5.6-Sol-Ultrafast, Gemma, Gemma-4-31B, Kimi K2.6, Llama, Llama / Qwen / GPT-OSS, QWEN 3.8 27B, Qwen, WSE-3Codex-Spark, GLM-4.7, GPT 5.6 Sol, GPT-5.6-Sol-Ultrafast, Gemma, Gemma-4-31B, Kimi K2.6, Kimi K3.1, Llama, Llama / Qwen / GPT-OSS, QWEN 3.8 27B, Qwen, WSE-3
source
Nov 6, 2026first seencapability
Tenstorrent Blackhole — New capabilities: Open-source Stack
153 tracked154 tracked · +Open-source Stack
source
Nov 6, 2026first seenversion
Google TPU Ironwood — Google TPU Ironwood moved to Ironwood (7th Generation)
7th generationIronwood (7th Generation)
source
Nov 6, 2026first seencapability
Google TPU Ironwood — New capabilities: AI Acceleration, Deep Learning
83 tracked85 tracked · +AI Acceleration, Deep Learning
source
Nov 5, 2026first seenversion
Groq LPU — Groq LPU moved to GroqCloud API
LPUGroqCloud API
source
Nov 5, 2026first seenversion
Google TPU Ironwood — Google TPU Ironwood moved to 7th generation
7th generation (Ironwood)7th generation
source
Nov 5, 2026first seenversion
Cerebras WSE-3 — Cerebras WSE-3 moved to WSE-3
CS-4WSE-3
source
Nov 4, 2026first seenversion
Tenstorrent Blackhole — Tenstorrent Blackhole moved to Blackhole
0.82.0-dev20261103Blackhole
source
Nov 4, 2026first seencapability
Tenstorrent Blackhole — New capabilities: Enterprise AI Server
152 tracked153 tracked · +Enterprise AI Server
source
Nov 4, 2026first seenversion
Google TPU Ironwood — Google TPU Ironwood moved to 7th generation (Ironwood)
Ironwood (7th Gen)7th generation (Ironwood)
source
Nov 4, 2026first seenversion
NVIDIA Blackwell — NVIDIA Blackwell moved to GB200 NVL72
BlackwellGB200 NVL72
source
Nov 4, 2026first seenversion
Cerebras WSE-3 — Cerebras WSE-3 moved to CS-4
WSE-3CS-4
source
Nov 3, 2026first seencapability
Intel Gaudi 3 — New capabilities: FP8/INT4 Inference
87 tracked88 tracked · +FP8/INT4 Inference
source
Nov 3, 2026first seenversion
Tenstorrent Blackhole — Tenstorrent Blackhole moved to 0.82.0-dev20261103
Blackhole0.82.0-dev20261103
source
Nov 3, 2026first seencapability
Tenstorrent Blackhole — New capabilities: AI hardware acceleration
151 tracked152 tracked · +AI hardware acceleration
source
Nov 3, 2026first seenversion
Huawei Ascend 910C — Huawei Ascend 910C moved to Ascend 910C
910CAscend 910C
source
Nov 3, 2026first seenversion
AWS Trainium2 — AWS Trainium2 moved to Trainium2
Trn2Trainium2
source
Nov 3, 2026first seenversion
Google TPU Ironwood — Google TPU Ironwood moved to Ironwood (7th Gen)
IronwoodIronwood (7th Gen)
source
Nov 3, 2026first seencapability
AMD Instinct MI355X — New capabilities: ROCm Core SDK 10.1.0
96 tracked97 tracked · +ROCm Core SDK 10.1.0
source
Nov 3, 2026first seencapability
NVIDIA Blackwell — New capabilities: Accelerated Computing
68 tracked69 tracked · +Accelerated Computing
source
Nov 2, 2026first seenmodel
Tenstorrent Blackhole — Tenstorrent Blackhole added Blackhole Tensix to its model stack
GLM-5.2, Gemma-2, Gemma4, Kimi-K3, Llama-3.3-70B, Qwen3-32B, Tensix cores, tt-metal / tt-nnBlackhole Tensix, GLM-5.2, Gemma-2, Gemma4, Kimi-K3, Llama-3.3-70B, Qwen3-32B, Tensix cores, tt-metal / tt-nn
source