Specs & Performance
Aether Platform offers diverse instance types ranging from general-purpose development to large-scale workloads.
Instance Types
Standard (CPU Optimized)
For Web development, Microservices, and lightweight data processing.
| Type | vCPU | RAM | Storage | Usage |
|---|---|---|---|---|
| Dev-S | 2 vCPU | 8 GB | 50 GB NVMe | Students, Learning, Static Sites |
| Dev-M | 4 vCPU | 16 GB | 100 GB NVMe | Frontend, Backend, Go/Rust |
| Dev-L | 8 vCPU | 32 GB | 200 GB NVMe | Microservices, Docker-Compose |
Upcoming: AI & GPU Support
As part of our roadmap, we plan to offer high-performance instances optimized for AI inference, fine-tuning, and data science.
| Type | GPU (Planned) | vRAM | vCPU | RAM | Usage |
|---|---|---|---|---|---|
| AI-S | NVIDIA T4 | 16 GB | 4 vCPU | 32 GB | Inference, Lightweight Training |
| AI-M | NVIDIA A10G | 24 GB | 8 vCPU | 64 GB | LoRA, Code Llama 7B |
| AI-L | NVIDIA A100 | 40 GB | 12 vCPU | 128 GB | Full Finetuning, 70B Models |
Region Support
Choose regions based on data residency requirements and latency.
- US-East (N. Virginia): Earliest access to latest instances.
- AP-Northeast (Tokyo): Compliance with Japanese data residency. Lowest latency for APAC.
- EU-Central (Frankfurt): GDPR compliant region.
Storage Performance
All workspaces are provisioned with high-throughput NVMe SSDs. We also support project-level NFS (Network File System) mounts, allowing multi-terabyte datasets to be shared across multiple workspaces.