Looking Server for AI Development (with gpu)

Comments

• georgedatacenter Member, Patron Provider

  February 1

  We have this available, but we’re not sure if it’s below what you need.

  AMD Ryzen 5950X
  16-Core, 32-Thread
  128GB DDR4
  2 x 2TB NVMe
  10TB Bandwidth
  Port 10 Gigabit
  1 IPv4
  IPMI

  $99 mon

  Order
• XenoxWork Member

  February 1

  @georgedatacenter said:
  We have this available, but we’re not sure if it’s below what you need.

  AMD Ryzen 5950X
  16-Core, 32-Thread
  128GB DDR4
  2 x 2TB NVMe
  10TB Bandwidth
  Port 10 Gigabit
  1 IPv4
  IPMI

  $99 mon

  Order

  need gpu for ai
• host_c Patron Provider, Top Host, Megathread Squad

  February 1

  @XenoxWork said:

  @georgedatacenter said:
  We have this available, but we’re not sure if it’s below what you need.

  AMD Ryzen 5950X
  16-Core, 32-Thread
  128GB DDR4
  2 x 2TB NVMe
  10TB Bandwidth
  Port 10 Gigabit
  1 IPv4
  IPMI

  $99 mon

  Order

  need gpu for ai

  You should specify that you need also a GPU with tensor cores for your project.
  Nvidia RTX3000 or better would be the bare minimum.
  Whatever Tesla P series will be almost useless for today standards.
• ComputeBox Member, Patron Provider

  February 1

  Hey there,
  we could offer u a 4000 ADA for 67,22€ / month (discount for the first billing period / net).

  Take a look at: https://computebox.de/#pricing

  Thanked by 1murathai
• NovaCloudHosting Member, Patron Provider

  February 1

  • i7-13700KF
  • 64GB RAM
  • 500GB NVMe SSD
  • RTX 5060Ti (16GB VRAM)
  • 1G Uplink
  • location: Germany, Frankfurt
    150€/Month excl. VAT + 20€ setup fee

  Upgrading RAM to 128GB, 10G Uplink or more NVMe Storage is possible. Just send me a DM.
• DataRecovery Member

  February 1

  @XenoxWork said:
  Training small/medium AI models
  Running LLMs
  Need Gpu %100
  min 16 gb Vram

  I don’t have strict hardware requirements.
  Just need something stable and reasonably priced.

  A couple of options with 24 GB VRAM and fully dedicated P40.
  Probably slower than bleeding edge GPUs, but should be enough for models like Deepseek-r1-0528-qwen3-8b, Llama-2-7B-Chat, Qwen3-4B-Thinking-2507, and Qwen3-4B-Instruct-2507-F16.

  GPU VDS:

  • 6 dedicated Xeon E5-2580 v2 cores // 30 GB RAM // 400 GB NVMe (ref)
  • Fully dedicated Nvidia Tesla P40 (PCI passthrough)
  • 100 TB transfer @ 10 Gbit/s (noref)

  Dedi (GPU is in the hardware options):

  • AMD EPYC 7K62 (48 cores, 96 threads) // 128 GB DDR4 ECC RAM (ref)
  • 3.2 TB Enterprise NVMe (supports extra NVMe, HDDs, RAM, and GPU)
  • 100 TB @ 10 Gbit/s ~or~ 1 Gbps unlimited, IPMI and console access (noref)

  / disable your VPN before placing an order /

  Thanked by 2host_c servarica_hani
• introserv_rd Member, Patron Provider

  February 4

  If you have an offer or can recommend a provider, feel free to message me.
  Thanks.

  Hi, we can offer AMD Ryzen 9900x based servers with Nvidia RTX 5070Ti in EU.

  Location: Netherlands (Naaldwijk)
  CPU: AMD Ryzen 9 9900X (12 Core 24 Threads 4.4/5.6GHz)
  Memory: 64GB REG ECC DDR5
  Storage: 2x 960GB Enterprise NVMe SSD (boot)
  GPU: Nvidia GeForce RTX 5070 TI-16G
  Port: 1 Gbit/s - Unmetered
  IPv4: Additional IPs available
  IP-KVM: IP-KVM by request

  Starts at €340.50/month Link to the configurator
• alexhost Member, Patron Provider

  February 4 edited February 4

  @XenoxWork said:
  Hi,

  I’m looking for a server that I can use for AI development.
  It will be used for:

  Training small/medium AI models

  Running LLMs

  Image and audio processing

  Need Gpu %100

  min 16 gb Vram

  I don’t have strict hardware requirements.
  Just need something stable and reasonably priced.

  If you have an offer or can recommend a provider, feel free to message me.

  Thanks.

  Hi.

  We have GPU Hosting Dedicated Servers for AI Models / LLMs.

  However sometimes we have it available and not available yet in the website. For that reason, we recommend you open a ticket support to know if have stock available.

  We have some GPU dedicated servers:
  https://alexhost.com/gpu-hosting/

  We have also Tesla GPU but is not under your requirements
  https://alexhost.com/gpu-hosting/dedicated-server-with-nvidia-gpu/

  Coupon Code: RED10
  RECURRING discount!

  Note: We recommend you to open a ticket support and we will notify you when more stock is available or if is available:
  https://bill.alexhost.com/tickets/new/

  Setup time: <24h max. Usually we deliver more faster.

  Best Regards,
  Alexhost
• murathai Member

  February 6

  @ComputeBox said:
  Hey there,
  we could offer u a 4000 ADA for 67,22€ / month (discount for the first billing period / net).

  Take a look at: https://computebox.de/#pricing

  ADA 4000 price looks very good.
• oomer Member

  February 8

  I'm testing a Bee VDS from servarica.com in Canada with pcie-passthrough Tesla P40, 6vcpu E5-2580v2 dedicated, 30GB Ram, 10Gbps, 400GB NVME, $79/mth .

  root@verily:~# nvidia-smi
  Sun Feb  8 03:04:54 2026       
  +---------------------------------------------------------------------------------------+
  | NVIDIA-SMI 535.288.01             Driver Version: 535.288.01   CUDA Version: 12.2     |
  |-----------------------------------------+----------------------+----------------------+
  | GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
  | Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
  |                                         |                      |               MIG M. |
  |=========================================+======================+======================|
  |   0  Tesla P40                      Off | 00000000:00:05.0 Off |                  Off |
  | N/A   21C    P8               9W / 250W |      0MiB / 24576MiB |      0%      Default |
  |                                         |                      |                  N/A |
  +-----------------------------------------+----------------------+----------------------+
  
  +---------------------------------------------------------------------------------------+
  | Processes:                                                                            |
  |  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
  |        ID   ID                                                             Usage      |
  |=======================================================================================|
  |  No running processes found                                                           |
  +---------------------------------------------------------------------------------------+
  

  Thanked by 1servarica_hani
• m3th3lesh Member

  February 8

  @DataRecovery said:

  @XenoxWork said:
  Training small/medium AI models
  Running LLMs
  Need Gpu %100
  min 16 gb Vram

  I don’t have strict hardware requirements.
  Just need something stable and reasonably priced.

  A couple of options with 24 GB VRAM and fully dedicated P40.
  Probably slower than bleeding edge GPUs, but should be enough for models like Deepseek-r1-0528-qwen3-8b, Llama-2-7B-Chat, Qwen3-4B-Thinking-2507, and Qwen3-4B-Instruct-2507-F16.

  GPU VDS:

  • 6 dedicated Xeon E5-2580 v2 cores // 30 GB RAM // 400 GB NVMe (ref)
  • Fully dedicated Nvidia Tesla P40 (PCI passthrough)
  • 100 TB transfer @ 10 Gbit/s (noref)

  Dedi (GPU is in the hardware options):

  • AMD EPYC 7K62 (48 cores, 96 threads) // 128 GB DDR4 ECC RAM (ref)
  • 3.2 TB Enterprise NVMe (supports extra NVMe, HDDs, RAM, and GPU)
  • 100 TB @ 10 Gbit/s ~or~ 1 Gbps unlimited, IPMI and console access (noref)

  / disable your VPN before placing an order /

  For running deepseek r-3? What is the option?

  Thanked by 1servarica_hani
• AndreiPerju Member, Host Rep

  February 8

  nVidia A40 is ok for you? We can make few configuration with that.
• Rackoona Member, Patron Provider

  February 8

  We at rackoona.com can offer you a4000 $99
• Nexeon Member, Patron Provider

  February 10

  @XenoxWork said:
  Hi,

  I’m looking for a server that I can use for AI development.
  It will be used for:

  Training small/medium AI models

  Running LLMs

  Image and audio processing

  Need Gpu %100

  min 16 gb Vram

  I don’t have strict hardware requirements.
  Just need something stable and reasonably priced.

  If you have an offer or can recommend a provider, feel free to message me.

  Thanks.

  We have RTX 5090 and RTX PRO 6000 in stock. Send me a message if interested.