Blog post

We’re excited to announce that our newest high-performance machine is now live and available to rent on Vast.ai.

This powerhouse comes packed with:

• 8× NVIDIA RTX 5090 GPUs – cutting-edge performance for training and inference
• 8 TB of NVMe storage – fast, reliable local disk for your datasets and checkpoints
• 386 GB of memory – plenty of headroom for large models
• 128 CPU threads – strong multi-threaded performance for preprocessing and hybrid workloads

Whether you’re fine-tuning LLMs, running massive inference pipelines, or experimenting with generative media, this server is designed to give you top-tier performance without the overhead of a full data center setup.

👉 You can rent it today directly on Vast.ai.