World's largest AI chip: 4T transistors, 125 petaflops on one wafer.
The WSE-3 is the largest AI chip ever built, measuring 46,225 mm² and containing 4 trillion transistors. It delivers 125 petaflops of AI compute through 900,000 AI-optimized cores — 19× more transistors and 28× more compute than the NVIDIA B200.
Purpose-built for ultra-fast AI training and inference with unmatched on-chip memory and bandwidth, single-device supercomputing simplicity, and fail-in-place redundant architecture for wafer-scale yield.
| Attribute | Wafer-Scale Engine 3 (WSE-3) | Blackwell, Nvidia B200 | LPU | |
|---|---|---|---|---|
| Provider |  |  |  | |
| Founded | 2016 | 1993 | 2016 | |
| Sells To | — | Consumers, Enterprise | Enterprise | |
| Pricing Model | — | Licensing, Product Sales, Subscription | Recurring | |
| Ownership | — | Public | Venture-backed |
Provider
Founded 2016
Sells To —
Pricing Model —
Ownership —
Provider
Founded 1993
Sells To Consumers, Enterprise
Pricing Model Licensing, Product Sales, Subscription
Ownership Public
Provider
Founded 2016
Sells To Enterprise
Pricing Model Recurring
Ownership Venture-backed
Prometheus is a memory-first AI server that collapses racks of infrastructure into one unit.
Corsair combines memory-centric hardware and software for low-latency AI inference in data centers.
RDU is SambaNova's Reconfigurable Dataflow Unit, the company's AI chip family for inference and training at scale.
Trainium is AWS's purpose-built AI accelerator chip family for training and inference of large generative models.
Tenstorrent PCIe accelerator card family for AI training and inference workloads.
Prometheus is a memory-first AI server that collapses racks of infrastructure into one unit.
Corsair combines memory-centric hardware and software for low-latency AI inference in data centers.
Sohu is a purpose-built inference chip designed for high-performance AI model serving.
SambaNova DataScale is a data-center AI computing platform.
RDU is SambaNova's Reconfigurable Dataflow Unit, the company's AI chip family for inference and training at scale.
Purpose-built AI supercomputers integrating GPUs, CPUs, networking, and software for enterprise training and inference.
TPU is Google's custom AI accelerator ASIC family delivered through Google Cloud TPU Pods for ML training and inference.
Trainium is AWS's purpose-built AI accelerator chip family for training and inference of large generative models.
Tenstorrent PCIe accelerator card family for AI training and inference workloads.
Intel Gaudi is a data-center AI accelerator product family.
