---
canonical: "https://www.softatlas.io/coreweave/products/nvidia-hopper"
name: "NVIDIA Hopper"
provider: "CoreWeave"
category: "Artificial Intelligence / AI Infrastructure / GPU Cloud Platforms"
review_count: 0
ai_powered: false
website: "https://www.coreweave.com/products/hgx-h100-h200"
updated: "2026-07-20"
---

# NVIDIA Hopper

By [CoreWeave](/coreweave.md)

NVIDIA Hopper delivers cutting-edge GPU technology for AI applications, supporting high-performance computing needs.

## Features
- Accelerate AI training with NVIDIA Transformer Engine
- Enhance LLM inference performance with HBM3e memory
- Optimize AI workloads with high-performance GPU compute
- Facilitate generative AI and LLMs with advanced memory bandwidth
- Enable large-scale AI model training with purpose-built infrastructure
- Support low-latency AI inference with high-performance compute
- Utilize GPU-to-GPU interconnect with NVIDIA NVLink
- Improve agent reliability through agentic AI and inference

## Built for
DEVELOPERS, OPERATIONS, EXECUTIVES

Source: https://www.softatlas.io/coreweave/products/nvidia-hopper