is now Verda
Learn more
AI Cloud
AI Cloud
GPUs
GPUs
Why Verda
Why Verda
Resources
Resources
About
About
.
Login
Sign up
AI Cloud
Instances
1x-8x GPUs with NVLink
Instant Clusters
16x-128x GPUs with InfiniBand™
Bare-metal Clusters
Configured to your specifications
Serverless Containers
Serverless GPUs with auto-scaling
Managed Endpoints
Inference API for SOTA models
GPUs
B300 SXM6
NEW
30 CPUs | 275 GB RAM | 262 GB VRAM
B200 SXM6
30 CPUs | 184 GB RAM | 180 GB VRAM
H200 SXM5
44 CPUs | 182 GB RAM | 141 GB VRAM
H100 SXM5
30 CPUs | 120 GB RAM | 80 GB VRAM
A100
22 CPUs | 120 GB RAM | 80 GB VRAM
See all
L40S, RTX 6000 ADA, and more
Why Verda
Resources
Blog
GPU benchmarks and R&D on Efficient AI
Trust Center
Security and compliance at Verda
Compute Credits
Earn compute credits for writing content
Docs
Technical documentation and tutorials
API
Control GPU resources via external code
Developer Community
Exchange knowledge with AI experts
About
Contact Us
Talk to our engineers or sales
Career
Join us to build Europe's first hyperscaler
Jobs
Explore the open positions
FAQ
Frequently asked questions
Company
Learn more about us as a company
Login
Sign up
Contact Us
Verda Blog
All
Customer stories
News
GPU comparison
Monthly digests
Blackwell
Guides
Benchmarks
Technical analyses
Legacy
Deutsch
English
Search
NEW
Blackwell
Dec 17, 2025
NVIDIA B200 & B300: GPU Architecture and Software Stack
