Dedicated Resources
Dedicated resources provide reserved compute capacity on LW AI's shared infrastructure. You get guaranteed throughput and lower latency without the complexity of managing your own deployment.
vs Private Deployment
| Aspect | Dedicated Resources | Private Deployment |
|---|---|---|
| Infrastructure | LW AI-managed | Your own |
| Setup time | Hours | Weeks |
| Data residency | LW AI's region | Your infrastructure |
| Cost model | Monthly subscription | Hardware + license |
| Customization | Limited | Full control |
| Best for | High-volume, predictable workloads | Strict compliance requirements |
Tiers
| Tier | GPU | RPM | TPM | Concurrent |
|---|---|---|---|---|
| Standard | 2x A100 | 200 | 1M | 30 |
| Pro | 4x A100 | 500 | 3M | 100 |
| Enterprise | 8x+ A100 | Custom | Custom | Custom |
Benefits
- Guaranteed throughput — Reserved capacity that's always available
- No rate limits — Your quota is guaranteed, not shared
- Lower latency — Dedicated GPU instances reduce queue time
- Priority routing — Requests skip the shared queue
Getting Started
- Contact enterprise sales
- Specify your throughput and model requirements
- We provision dedicated resources within 24 hours
- Use the same API endpoint with your dedicated API key
Related
- Private Deployment — Full infrastructure isolation
- Rate Limits — Shared tier limits