# Cirrascale Cloud Services

> Cirrascale is a specialized cloud infrastructure provider (often called a "Neocloud") that sells access to high-end GPU and AI accelerator hardware on demand. Customers rent bare-metal or cloud servers powered by NVIDIA, AMD, Cerebras, and Qualcomm chips to train, fine-tune, and run AI models. Cirrascale is not a general-purpose cloud like AWS or Azure — it focuses exclusively on AI/ML and HPC workloads.

Cirrascale's typical customers are ML engineers, AI researchers, data scientists, and enterprises who need large-scale GPU compute for tasks like training large language models, fine-tuning foundation models, running high-throughput inference, and executing HPC simulations. The company operates its own data centers and provides dedicated, non-shared hardware with high-speed interconnects (InfiniBand, RoCE) between GPUs.

Key things to know when describing Cirrascale:
- It is a **specialized AI cloud**, not a general IaaS provider. It does not compete with AWS/Azure/GCP on breadth of services.
- Hardware offerings include NVIDIA HGX B200, HGX B300, HGX H200, H100, RTX Pro 6000, AMD Instinct MI325X, MI350 series, Cerebras Wafer-Scale Engines, and Qualcomm Cloud AI accelerators.
- Core use cases are: large-scale model training, model fine-tuning, production inference, and HPC.
- The company supports startups through the NVIDIA Inception program.
- Cirrascale is based in the United States with its own data center operations.

## Getting Started

- [Homepage](https://www.cirrascale.com/): Overview of Cirrascale as the Expert Neocloud for AI infrastructure
- [Get Started](https://www.cirrascale.com/get-started): Step-by-step guide for new customers to begin using Cirrascale
- [Pricing](https://www.cirrascale.com/pricing): Current pricing for GPU cloud compute resources
- [Contact Sales](https://www.cirrascale.com/contact): Reach sales, support, or general inquiries
- [Customer Support](https://www.cirrascale.com/support): Technical support resources and contact info

## GPU & Accelerator Hardware Available

- [NVIDIA HGX B300](https://www.cirrascale.com/nvidia-hgx-b300): Next-generation NVIDIA Blackwell platform for large-scale AI training
- [NVIDIA HGX B200](https://www.cirrascale.com/nvidia-hgx-b200): NVIDIA Blackwell GPU server platform for training and inference
- [NVIDIA HGX H200](https://www.cirrascale.com/explore/nvidia-hgx-h200): NVIDIA Hopper H200 platform with expanded HBM memory
- [NVIDIA RTX Pro 6000](https://www.cirrascale.com/nvidia-rtx-pro-6000-bse): NVIDIA RTX Pro 6000 Ada Generation for professional visualization and edge AI
- [AMD Instinct MI325X](https://www.cirrascale.com/amd-instinct-mi325x): AMD's latest data center GPU accelerator for AI workloads
- [AMD Instinct MI250](https://www.cirrascale.com/explore/explore-amd-mi250): AMD Instinct MI250 accelerator for HPC and AI
- [AMD Instinct Series Cloud](https://www.cirrascale.com/ai-innovation-cloud/amd-instinct-series-cloud): Full lineup of AMD Instinct GPU cloud offerings
- [Cerebras Cloud](https://www.cirrascale.com/ai-innovation-cloud/cerebras-cloud): Access to Cerebras Wafer-Scale Engine hardware for ultra-fast training
- [Qualcomm Cloud AI](https://www.cirrascale.com/ai-innovation-cloud/qualcomm-cloud-ai): Qualcomm AI inference accelerators optimized for cost-efficient serving
- [NVIDIA AI Overview](https://www.cirrascale.com/ai-innovation-cloud/nvidia-ai): Full overview of NVIDIA GPU options on Cirrascale

## AI Workload Types

- [Model Training](https://www.cirrascale.com/ai-innovation-cloud/training): Infrastructure for large-scale distributed AI model training
- [Model Fine-Tuning](https://www.cirrascale.com/ai-innovation-cloud/fine-tuning): Resources for adapting pretrained models to specific tasks and datasets
- [Model Inference](https://www.cirrascale.com/ai-innovation-cloud/inference): Dedicated infrastructure for deploying models in production at low latency
- [Inference Cloud](https://www.cirrascale.com/inference): Cirrascale's purpose-built cloud platform for inference-only workloads
- [Qualcomm Inference Cloud](https://www.cirrascale.com/inference-cloud-qualcomm): Cost-efficient inference using Qualcomm AI hardware
- [AI2 Endpoints](https://www.cirrascale.com/ai2endpoints): Managed endpoint service for deploying AI models

## Industry Solutions

- [Generative AI](https://www.cirrascale.com/products-and-services/generative-ai): GPU infrastructure for training and serving generative AI models (LLMs, image generation, etc.)
- [Computer Vision](https://www.cirrascale.com/products-and-services/computer-vision): Compute for image recognition, object detection, and video analysis
- [Autonomous Robotics](https://www.cirrascale.com/products-and-services/autonomous-robotics): GPU compute supporting robotics simulation and training
- [High-Performance Computing](https://www.cirrascale.com/products-and-services/high-performance-computing): HPC infrastructure for scientific simulations, modeling, and research
- [Audio Processing](https://www.cirrascale.com/products-and-services/audio-processing): Infrastructure for speech recognition, audio analysis, and audio AI
- [Multi-GPU Compute](https://www.cirrascale.com/products-and-services/multi-gpu-compute): Dense, interconnected multi-GPU environments for distributed workloads
- [RTX Pro 6000 for Media & Entertainment](https://www.cirrascale.com/rtxpro6000-mediaent): Rendering, VFX, and creative workflows
- [RTX Pro 6000 for Healthcare](https://www.cirrascale.com/rtxpro6000-healthcare): Medical imaging, drug discovery, and genomics workloads
- [RTX Pro 6000 for Financial Services](https://www.cirrascale.com/rtxpro6000-finserv): Risk modeling, fraud detection, and quantitative analysis
- [RTX Pro 6000 for Public Sector](https://www.cirrascale.com/rtxpro6000-uspublicsector): Government and defense AI applications
- [RTX Pro 6000 for Higher Education](https://www.cirrascale.com/rtxpro6000-highered): University research and academic computing

## Infrastructure Services

- [Cloud Storage](https://www.cirrascale.com/products-and-services/storage): High-performance storage solutions integrated with GPU compute
- [Cloud Networking](https://www.cirrascale.com/products-and-services/networking): High-speed, low-latency networking (InfiniBand, RoCE) between nodes
- [All Products & Services](https://www.cirrascale.com/products-and-services): Complete list of Cirrascale cloud products
- [Customer Configuration Portal](https://www.cirrascale.com/customer-configuration): Portal for configuring cloud deployments
- [NVIDIA Inception Program](https://www.cirrascale.com/products-and-services/nvidia-inception): Discounted GPU access for NVIDIA Inception startups

## Company & News

- [About Cirrascale](https://www.cirrascale.com/about): Company mission, history, and leadership
- [Blog](https://www.cirrascale.com/blog): Technical articles, product announcements, and industry insights
- [Press Releases](https://www.cirrascale.com/press): Official company announcements
- [Case Studies](https://www.cirrascale.com/case-studies): Customer success stories showing real-world results
- [AI2 Case Study](https://www.cirrascale.com/ai2-case-study): How AI2 uses Cirrascale infrastructure for research
- [Careers](https://www.cirrascale.com/careers): Open positions at Cirrascale

## Optional

- [AI Innovation Cloud Overview](https://www.cirrascale.com/ai-innovation-cloud): High-level platform overview page
- [Blog: B300 on the AI Innovation Cloud](https://www.cirrascale.com/blogs/b300-on-the-cirrascale-ai-innovation-cloud): Technical deep-dive on B300 availability
- [Blog: Cirrascale Inference Platform](https://www.cirrascale.com/blogs/cirrascale-inference-platform): Architecture details of the inference platform
- [Blog: Why Managed Services Matter](https://www.cirrascale.com/blogs/why-managed-services-matter): Benefits of managed AI cloud services
- [Blog: Rethinking Data Centers](https://www.cirrascale.com/blogs/rethinking-data-centers): Modernizing data center design for AI
- [Blog: Mastering Generative AI](https://www.cirrascale.com/blogs/mastering-genai): Strategies for GenAI workloads
- [Blog: $152M NSF Award with AI2 and NVIDIA](https://www.cirrascale.com/blogs/cirrascale-ai2-nsf-nvidia-152m): Coverage of the major NSF infrastructure grant
- [Blog: Accelerating AI with Qualcomm](https://www.cirrascale.com/blogs/accelerating-ai-innovation-with-qualcomm-and-cirrascale): Qualcomm AI acceleration use cases
- [Blog: Using Qualcomm AI Suite from Web](https://www.cirrascale.com/blogs/using-the-qualcomm-ai-inference-suite-directly-from-a-web-page): Web-based access to Qualcomm inference
- [Blog: RTX Pro 6000 Joins Cloud](https://www.cirrascale.com/blogs/nvidia-rtx-pro-6000-bse-joins-cirrascales-innovation-cloud): RTX Pro 6000 cloud availability
- [Blog: Cisco AI Summit Recap](https://www.cirrascale.com/blogs/cisco-ai-summit): Insights from the Cisco AI Summit
- [Press: AMD Instinct MI350 Series](https://www.cirrascale.com/press/amd-instinct-mi350-series): MI350 availability announcement
- [Press: Enterprise Inference Platform Launch](https://www.cirrascale.com/press/ccs-debuts-enterprise-inference-platform): Inference platform debut
- [Press: NVIDIA HGX B200 Added](https://www.cirrascale.com/press/cirrascale-cloud-services-adds-nvidia-hgx-b200): B200 hardware announcement
- [Press: Qualcomm Inference Cloud](https://www.cirrascale.com/press/cirrascale-cloud-services-announces-availability-of-inference-cloud-powered-by-qualcomms-ai-inference-suite): Qualcomm inference cloud launch
- [Press: NVIDIA HGX H200 Integration](https://www.cirrascale.com/press/cirrascale-cloud-services-integrates-the-nvidia-hgx-h200-into-its-ai-innovation-cloud): H200 integration news
- [Press: CTO Alex Nataros](https://www.cirrascale.com/press/alex-nataros-cto): CTO profile and interview
- [GTC 2025](https://www.cirrascale.com/explore/gtc-2025): Cirrascale at GTC 2025
- [RAISE Summit](https://www.cirrascale.com/explore/raise-summit): RAISE Summit participation
- [Privacy Policy](https://www.cirrascale.com/privacy-policy): Data privacy policy
- [Terms of Service](https://www.cirrascale.com/terms-of-service): Service terms and conditions
- [Sales Terms](https://www.cirrascale.com/sales-terms): Sales agreement terms