Trooper.AI delivers NVIDIA RTX 3090 24 GB GPU servers with full power and zero overhead. Whether youβre training large models, running complex AI inference, or fine-tuning LLMs β our next-generation RTX 3090 instances are ready for production use, hosted securely in the EU.
Our RTX 3090 24 GB instances offer:
We love AI to serve you with the best AI ever:



| GPU | VRAM | SDXL (16 imgs)* | LLM 8b Token/s** |
|---|---|---|---|
| RTX 4070S Ti | 16 GB | Ada | 2:17 min |
| RTX 3090 | 24 GB | Ampere | 2:24 min |
| Tesla V100 | 16/32 GB | Volta | 2:36 min |
| RTX A4000 | 16 GB | Ampere | 3:37 min |
** Speed benchmark conducted using Automatic1111 v1.6.0 with the following settings:
Model: sd_xl_base.1.0,
Prompt: βcute Maltese puppy on green grassβ (no negative prompt),
Sampler: DPM++ 2M Karras,
No Refiner, No Hires Fix,
CFG Scale: 7, Steps: 30,
Resolution: 1024Γ1024,
Batch Count: 4, Batch Size: 4,
Random Seed, PNG Images.
** LLM speed benchmark conducted using a default OpenWebUI installation without any modifications.
Context length: 2048 (default).
No system prompt.
Query prompt: βName the 5 biggest similarities between a wild tiger and a domestic cat.β
Model (small): llama3.1:8b-instruct-q8_0
Best of 3 runs, measured in response tokens per second β higher values indicate better performance.
Explore our RTX 3090 24 GB server options and pricing. Servers hosted with a β.deβ domain are located in Germany within ISO/IEC 27001 certified data centers.
| Blib Name | GPU | CPU RAM | Cores | NVMe | Price | Promo Price* |
|---|---|---|---|---|---|---|
| sparbox.xs1 | 1Γ RTX 3090 24 GB | 28 GB | 2 | 180 GB | β¬0.17/h or β¬100/m | |
| sparbox.s1 | 1Γ RTX 3090 24 GB | 32 GB | 4 | 180 GB | β¬0.18/h or β¬103/m | |
| sparbox.ms1 | 1Γ RTX 3090 24 GB | 56 GB | 4 | 180 GB | β¬0.19/h or β¬105/m | |
| sparbox.xs1.de | 1Γ RTX 3090 24 GB | 26 GB | 3 | 180 GB | β¬0.19/h or β¬108/m | |
| sparbox.sm1 | 1Γ RTX 3090 24 GB | 32 GB | 4 | 600 GB | β¬0.19/h or β¬110/m | |
| sparbox.m1 | 1Γ RTX 3090 24 GB | 56 GB | 4 | 600 GB | β¬0.20/h or β¬115/m | |
| sparbox.l1 | 1Γ RTX 3090 24 GB | 56 GB | 8 | 900 GB | β¬0.22/h or β¬123/m | |
| sparbox.l1.de | 1Γ RTX 3090 24 GB | 56 GB | 8 | 900 GB | β¬0.23/h or β¬135/m | |
| sparbox.m2 | 2Γ RTX 3090 24 GB | 58 GB | 8 | 680 GB | β¬0.35/h or β¬200/m | |
| sparbox.l2 | 2Γ RTX 3090 24 GB | 96 GB | 12 | 1200 GB | β¬0.38/h or β¬220/m | |
| sparbox.m2.de | 2Γ RTX 3090 24 GB | 58 GB | 12 | 900 GB | β¬0.39/h or β¬223/m | |
| sparbox.l2.de | 2Γ RTX 3090 24 GB | 96 GB | 12 | 1200 GB | β¬0.41/h or β¬235/m | |
| sparbox.m4 | 4Γ RTX 3090 24 GB | 128 GB | 16 | 900 GB | β¬0.68/h or β¬388/m | |
| sparbox.l4 | 4Γ RTX 3090 24 GB | 150 GB | 24 | 1600 GB | β¬0.72/h or β¬413/m |
* Weekly and monthly subscriptions offer reduced rates. Promotional pricing includes a 100% budget bonus, valid for this month.
We meticulously maintain our servers, exemplified by builds like the RTX 3090 24GB. This allows you to focus on your AI projects β development or production β without hardware concerns. As a boutique hyperscaler in Europe, we provide dedicated attention to detail and performance.
Rent your own GPU server today and start building amazing AI applications! Trooper.AI GPU servers are built from purely upcycled high-end tech from the last years, designed to provide you with the best performance, security, and reliability for all your AI needs.
EU location Β· High privacy Β· Great performance Β· Best support