GB300
GB300 is a NVIDIA CPU and GPU superchip based on the Grace CPU + 2x B300 architecture, first released in 2025.
Built from BenchGecko data as of April 14, 2026 · updates when the data changes
GB300 is a NVIDIA CPU and GPU superchip based on the Grace CPU + 2x B300 architecture, first released in 2025.
Basic
The NVIDIA Grace Blackwell Ultra GB300 is an AI CPU and GPU superchip made by NVIDIA, codenamed Grace Blackwell Ultra. Peak throughput is 5,000 TFLOPS at FP16/BF16 and 10,000 TFLOPS at FP8. It carries 576 GB of HBM3e with 16 TB/s of memory bandwidth. Rated power (TDP) is 3,200 W. It is built on N4P at TSMC.
Deep
In the BenchGecko dataset it sits in the frontier tier: current-generation flagship silicon shipping at scale to hyperscalers. It follows the GB200. Disclosed buyers include xAI. Specs come from manufacturer datasheets; the spec card on this page shows the dataset values and their as-of date.
Expert
That is 1.56 FP16 TFLOPS per watt. Memory bandwidth per unit of compute is 3.2 GB/s per FP16 TFLOP: a higher ratio helps memory-bound inference, while peak TFLOPS matter more for compute-bound training. Fabrication: Taiwan. Real throughput depends on the model, precision, batch size and software stack, so compare tokens per second per dollar for your own workload.
Depending on why you're here
- ·GB300: 5,000 TFLOPS FP16/BF16 · 10,000 TFLOPS FP8
- ·Process: N4P at TSMC
- ·Tier: frontier
- ·576 GB of HBM3e per chip
- ·No public cloud listing for GB300 in the dataset
- ·Compare it with other chips on /hardware/gb300
- ·Disclosed buyers: xAI
- ·NVIDIA · fabbed at TSMC
- ·Supply tightness is tracked on /hardware
- ·GB300 is a chip built to run AI, made by NVIDIA
- ·It has 576 GB of fast memory so large models fit
- ·Companies buy or rent these to train and serve AI models
Frequently Asked Questions
Read the primary sources
- Grace Blackwell Ultrawww.nvidia.com