TensorDock
Best for Budget Compute, Side Projects, Decentralized Rendering
from$0.10/ hr★ 8.9/10
SF Compute
Available NowBest for Funded AI Startups, Y Combinator Companies, LLM Foundation Models
About SF Compute
Born out of the extreme GPU shortages of the generative AI boom, SF Compute positions itself as the premier compute partner for high-growth AI startups. Rather than fighting for spot instances on AWS, funded companies can rely on SF Compute to provide dedicated, uninterrupted H100 GPU availability.
Founder-Friendly ML Compute
SF Compute is famous for its white-glove, founder-to-founder support model. Their premium AI startup clusters are architected specifically to reduce training bottlenecks for Large Language Models, featuring massive NVLink interconnects and blistering fast parallel storage systems.
Pros & Cons
Pros
- Secures uninterrupted H100 GPU availability for highly vetted startups
- Founder-friendly ML compute platform with incredibly fast customer support
- Clusters are highly optimized for training large scale generative models
- Zero complex egress fee structures
Cons
- Capacity is heavily gated; you generally have to apply or be a funded startup to gain access
- Does not offer cheap consumer GPUs or free tiers for hobbyists
- Geographic footprint is currently limited
Ideal Use Cases
enterprise-aiLLM Training
| GPU Models | H100, A100 |
| GPU Types | A100, H100 |
| Headquarters | San Francisco, CA, USA |
| Founded | 2023 |
| Availability | Available Now |
| Website | sfcompute.com ↗ |
$2.00/ hour (starting)—$4.50/ hr (max)
💡 Pricing note: Rates shown are indicative. Final pricing depends on GPU model, reservation type (spot vs. on-demand), contract length, and region. Get an exact quote →
US
Starting from
$2.00/hr
Up to $4.50/hr
Frequently Asked Questions
SF Compute GPU cloud pricing starts from $2.00/hr depending on GPU type, reservation model (on-demand vs. spot vs. reserved), and region. Use the quote form to get exact pricing for your specific workload.
SF Compute offers H100, A100 GPU instances. Availability varies by region and configuration. Contact the provider through ComputeStacker for current availability.
SF Compute operates data centers in US West. Choosing a region close to your users minimises latency and can help with data residency compliance requirements.
Use the "Get a Quote" button on this page to submit your GPU requirements. ComputeStacker will forward your request to SF Compute and other matching providers. You'll receive proposals within 24 hours — no commitment required.
SF Compute offers high-performance GPU infrastructure suitable for large language model training and fine-tuning workloads. For large-scale distributed training, check the Specs tab for NVLink and InfiniBand interconnect availability.
Explore More Options
Similar Providers
Compare Head-to-Head
Alternatives to SF Compute
Genesis Cloud
Best for GDPR-Compliant AI, European Data Sovereignty, Image Generation, Fine-Tuning
from$0.30/ hr★ 8.4/10
Civo
Best for Kubernetes-native AI applications, Developer deployments
from$0.20/ hr★ 8.8/10