Press & Media Kit
Resources for journalists, analysts, and partners covering VoltageGPU.
Company Boilerplate
VoltageGPU is a GPU cloud platform built on the Bittensor decentralized AI network. It aggregates GPU capacity from 4 specialized subnets — Lium (SN51) for compute, Chutes (SN64) for inference, Gradients (SN56) for fine-tuning, and Targon (SN4) for confidential compute — to deliver GPU pods, AI inference API, managed fine-tuning, and Intel TDX encrypted enclaves at competitive prices.
Founded in 2024 by Julien Aubry in Lyon, France, VoltageGPU serves developers, startups, and enterprise teams worldwide with per-second billing, 140+ pre-deployed AI models, and an OpenAI-compatible API.
Quick Facts
Products
GPU Compute
On-demand GPU pods with SSH access. RTX 4090 from $0.37/hr, H100 from $2.77/hr. Pre-installed CUDA, PyTorch, TensorFlow.
AI Inference API
140+ models via OpenAI-compatible API. DeepSeek R1, Qwen3, Llama, FLUX. From $0.02/M tokens.
Fine-Tuning
Managed training on decentralized GPUs. Text SFT, Chat, DPO, Image LoRA. From $9.25/hr.
Confidential Compute
Intel TDX encrypted GPU enclaves via Bittensor SN4. B200, H200, H100. Zero-knowledge to host.
Pricing Comparison (Citable)
| GPU | VoltageGPU | AWS | Savings |
|---|---|---|---|
| RTX 4090 (24GB) | $0.37/hr | ~$1.50/hr* | 75% |
| A100 80GB | $2.02/hr | $4.10/hr | 51% |
| H100 80GB | $2.77/hr | $12.36/hr | 78% |
| H200 141GB | $4.07/hr | $8.49/hr | 52% |
| B200 192GB | $7.50/hr | N/A | — |
*AWS pricing as of March 2026. RTX 4090 not available on AWS; comparison uses equivalent tier. VoltageGPU prices are live from decentralized network. All prices per-GPU, on-demand.
Brand Assets
Brand Colors
Violet #7c3aed · Green #4ade80 · Dark #050508
Typography
Inter, weight 400-800