Shadow offers European AI inference with listed per-minute GPU fees
In brief
French cloud-gaming provider Shadow has launched Blade Models and Blade AI Fabric. According to the company, Blade Models gives access to main open-weight models for text, images and audio through a web interface and an OpenAI-compatible API, priced per million tokens. Blade AI Fabric runs Python scripts on Shadow's servers, billed by the minute for GPU and CPU.
Source : LeMagIT - « Shadow convertit ses PC gamers en cloud en services d’IA, utilisables le jour »
What it changes for an SME
For your SME, this adds a European inference option with GPU prices listed by the hour and billed by the minute: according to Shadow, 40 euro cents an hour for an Nvidia RTX4500 with 20 GB of RAM, half that for an RTX2000 with 16 GB, and 5 euro cents an hour per CPU core. According to CEO Boris Lecoeur, US hyperscaler token and GPU prices are hard to read and you cannot be sure which law applies to work sent to those services. You are limited to infrastructure on your continent, and the consoles include a scheduler for cheaper hours, usage tracking and a budget cap. Shadow has not stated the extra night-time charge.
What you can do
- Compare Blade Models' price per million tokens with your current AI API volumes.
- Check whether your Python scripts can run on Blade AI Fabric under a budget cap.
- Confirm that processing would stay in Shadow's European data centres before sending company data.
Let's talk about the time you could save.
In 20 minutes, we find where AI could save you time or money. Free, no commitment.