NVIDIAB300
NVIDIA DGX B300
- GPU
- 8 x NVIDIA B300 SXM
- System memory
- 2000GB (DDR5 ECC REG)
- CPU
- 2x Intel Xeon 6776P (64c/128t, 2.3GHz-3.9GHz, 350W)
Server catalog
Twenty servers. Four GPU families.
Each listing is one physical machine with eight NVIDIA GPUs. Read the full spec sheet, then rent it by the hour in euros.
On demand€3.00 to €6.00/GPU·hr·4 builders
Showing 20 of 20 servers
Showing 20 of 20 servers/4 GPU families/4 buildersOn-demand · stop any time
NVIDIA B300
Blackwell Ultra8 listingsThe newest silicon we rent, with the most memory per GPU. Built for models that should not be cut into pieces.
- per GPU
- 288 GB HBM3e
- per server
- 2,304 GB
- from
- €48.00/hr
- NVIDIA · B300€48.00/server·hr€6.00/GPU·hr2000GB (DDR5 ECC REG)€48.00/server·hr€6.00/GPU·hr
- GIGABYTE · B300 · 8U€48.00/server·hr€6.00/GPU·hr6144GB (DDR5 ECC REG)€48.00/server·hr€6.00/GPU·hr
- GIGABYTE · B300 · 8U€48.00/server·hr€6.00/GPU·hr6144GB (DDR5 ECC REG)€48.00/server·hr€6.00/GPU·hr
- GIGABYTE G894-SD3-AAX78192GB RAMGIGABYTE · B300 · 8U€48.00/server·hr€6.00/GPU·hr8192GB (DDR5 ECC REG)€48.00/server·hr€6.00/GPU·hr
- Supermicro SYS-822GS-NB3RT HGX32 x 96GB RAMSupermicro · B300 · 8U€48.00/server·hr€6.00/GPU·hr32 x 96GB DDR5-6400 ECC RDIMM€48.00/server·hr€6.00/GPU·hr
- ASUS · B300€48.00/server·hr€6.00/GPU·hr32kh 128GB DDR5 RDIMM 6400MHz (Support up to 32 DIMM 4TB DDR5)€48.00/server·hr€6.00/GPU·hr
- GIGABYTE G894-SD3-AAX716x 128GB RAMGIGABYTE · B300 · 8U€48.00/server·hr€6.00/GPU·hr16x 128GB RDIMM 6400MHz (Support up to 32 DIMM DDR5 up to 8192 Gb)€48.00/server·hr€6.00/GPU·hr
- Supermicro SYS-822GS-NB3RT HGX24 x 96GB RAMSupermicro · B300 · 8U€48.00/server·hr€6.00/GPU·hr24 x 96GB DDR5-6400 ECC RDIMM€48.00/server·hr€6.00/GPU·hr
NVIDIA B200
Blackwell8 listingsBlackwell with room for large models. A strong default for training runs and high-traffic endpoints.
- per GPU
- 180 GB HBM3e
- per server
- 1,440 GB
- from
- €40.00/hr
- ASUS · B200€40.00/server·hr€5.00/GPU·hr8192GB (DDR5 ECC REG)€40.00/server·hr€5.00/GPU·hr
- NVIDIA · B200 · 10U€40.00/server·hr€5.00/GPU·hr2000GB (DDR5 ECC REG)€40.00/server·hr€5.00/GPU·hr
- GIGABYTE · B200 · 8U€40.00/server·hr€5.00/GPU·hr6144GB (DDR5 ECC REG)€40.00/server·hr€5.00/GPU·hr
- Supermicro · B200€40.00/server·hr€5.00/GPU·hr6144GB (DDR5 ECC REG)€40.00/server·hr€5.00/GPU·hr
- Supermicro · B200€40.00/server·hr€5.00/GPU·hr8192GB (DDR5 ECC REG)€40.00/server·hr€5.00/GPU·hr
- GIGABYTE · B200 · 8U€40.00/server·hr€5.00/GPU·hr6144GB (DDR5 ECC REG)€40.00/server·hr€5.00/GPU·hr
- GIGABYTE · B200 · 8U€40.00/server·hr€5.00/GPU·hr4096GB (DDR5 ECC REG)€40.00/server·hr€5.00/GPU·hr
- GIGABYTE · B200 · 8U€40.00/server·hr€5.00/GPU·hr8192GB (DDR5 ECC REG)€40.00/server·hr€5.00/GPU·hr
NVIDIA H200
Hopper2 listingsThe Hopper software stack you already know, with far more memory per GPU than the H100.
- per GPU
- 141 GB HBM3e
- per server
- 1,128 GB
- from
- €32.00/hr
- GIGABYTE · H200€32.00/server·hr€4.00/GPU·hr6144GB (DDR5 ECC REG)€32.00/server·hr€4.00/GPU·hr
- NVIDIA · H200 · 8U€32.00/server·hr€4.00/GPU·hr2000GB (DDR5 ECC REG)€32.00/server·hr€4.00/GPU·hr
NVIDIA H100
Hopper2 listingsProven Hopper nodes at the lowest hourly rate in the fleet. Ideal for steady inference and fine-tuning.
- per GPU
- 80 GB HBM3
- per server
- 640 GB
- from
- €24.00/hr
- ASUS · H100 · 7U€24.00/server·hr€3.00/GPU·hr6144GB (DDR5 ECC REG)€24.00/server·hr€3.00/GPU·hr
- NVIDIA · H100€24.00/server·hr€3.00/GPU·hr32x 64GB DDR5 4800MHz€24.00/server·hr€3.00/GPU·hr
NVIDIA B300
Blackwell Ultra8 listingsThe newest silicon we rent, with the most memory per GPU. Built for models that should not be cut into pieces.
- per GPU
- 288 GB HBM3e
- per server
- 2,304 GB
- from
- €48.00/hr
GIGABYTE8UB300
GIGABYTE G894-ZD3-AAX7
- GPU
- 8 x NVIDIA B300 SXM
- System memory
- 6144GB (DDR5 ECC REG)
- CPU
- Configurable on request
GIGABYTE8UB300
GIGABYTE G894-AD3-AAX7
- GPU
- 8 x NVIDIA B300 SXM
- System memory
- 6144GB (DDR5 ECC REG)
- CPU
- Configurable on request
- GPU
- 8 x NVIDIA B300 SXM
- System memory
- 8192GB (DDR5 ECC REG)
- CPU
- Configurable on request
- GPU
- 1 x NVIDIA HGX B300 NVL8 (8 GPU)
- System memory
- 32 x 96GB DDR5-6400 ECC RDIMM
- CPU
- 2 x Intel Xeon 6768P (64C 336MB Cache 2.40 GHz)
ASUSB300
ASUS XA NB3I-E12 10NVMe
- GPU
- 1 x NVIDIA HGX B300 8-GPU server
- System memory
- 32kh 128GB DDR5 RDIMM 6400MHz (Support up to 32 DIMM 4TB DDR5)
- CPU
- 2x Intel Xeon 6 6767P (64C 336M Cache 2.40 GHz)
- GPU
- B300 HGX
- System memory
- 16x 128GB RDIMM 6400MHz (Support up to 32 DIMM DDR5 up to 8192 Gb)
- CPU
- 2x Intel Xeon 6767P
- GPU
- 1 x NVIDIA HGX B300 NVL8 (8 GPU)
- System memory
- 24 x 96GB DDR5-6400 ECC RDIMM
- CPU
- 2 x Intel Xeon 6768P (64C 336MB Cache 2.40 GHz)
NVIDIA B200
Blackwell8 listingsBlackwell with room for large models. A strong default for training runs and high-traffic endpoints.
- per GPU
- 180 GB HBM3e
- per server
- 1,440 GB
- from
- €40.00/hr
ASUSB200
ASUS ESC NB8-E11
- GPU
- 8 x NVIDIA B200 SXM
- System memory
- 8192GB (DDR5 ECC REG)
- CPU
- Configurable on request
NVIDIA10UB200
NVIDIA DGX B200
- GPU
- 8 x NVIDIA B200 SXM
- System memory
- 2000GB (DDR5 ECC REG)
- CPU
- 2x Intel Xeon Platinum 8570 (56c/112t, 2.1GHz-4GHz, 350W)
GIGABYTE8UB200
GIGABYTE G894-AD1-AAX5
- GPU
- 8 x NVIDIA B200 SXM
- System memory
- 6144GB (DDR5 ECC REG)
- CPU
- Configurable on request
SupermicroB200
Supermicro AS-A126GS-TNBR
- GPU
- 8 x NVIDIA B200 SXM
- System memory
- 6144GB (DDR5 ECC REG)
- CPU
- Configurable on request
SupermicroB200
Supermicro SYS-A21GE-NBRT
- GPU
- 8 x NVIDIA B200 SXM
- System memory
- 8192GB (DDR5 ECC REG)
- CPU
- Configurable on request
GIGABYTE8UB200
GIGABYTE G893-ZD1-AAX5
- GPU
- 8 x NVIDIA B200 SXM
- System memory
- 6144GB (DDR5 ECC REG)
- CPU
- Configurable on request
GIGABYTE8UB200
GIGABYTE G894-SD1-AAX5
- GPU
- 8 x NVIDIA B200 SXM
- System memory
- 4096GB (DDR5 ECC REG)
- CPU
- Configurable on request
GIGABYTE8UB200
GIGABYTE G893-SD1-AAX5
- GPU
- 8 x NVIDIA B200 SXM
- System memory
- 8192GB (DDR5 ECC REG)
- CPU
- Configurable on request
NVIDIA H200
Hopper2 listingsThe Hopper software stack you already know, with far more memory per GPU than the H100.
- per GPU
- 141 GB HBM3e
- per server
- 1,128 GB
- from
- €32.00/hr
GIGABYTEH200
GIGABYTE G593-ZD1-AAX3
- GPU
- 8 x NVIDIA H200
- System memory
- 6144GB (DDR5 ECC REG)
- CPU
- Configurable on request
NVIDIA8UH200
NVIDIA DGX H200
- GPU
- 8 x NVIDIA H200
- System memory
- 2000GB (DDR5 ECC REG)
- CPU
- 2x Intel Xeon Platinum 8480C (56c/112t, 2GHz-3.8GHz, 350W)
NVIDIA H100
Hopper2 listingsProven Hopper nodes at the lowest hourly rate in the fleet. Ideal for steady inference and fine-tuning.
- per GPU
- 80 GB HBM3
- per server
- 640 GB
- from
- €24.00/hr
ASUS7UH100
ASUS ESC N8A-E12
- GPU
- 8 x NVIDIA H100
- System memory
- 6144GB (DDR5 ECC REG)
- CPU
- Configurable on request
NVIDIAH100
NVIDIA DGX H800
- GPU
- NVIDIA DELTA-NEXT Vulcan SXM5 8xNVIDIA H100 640G
- System memory
- 32x 64GB DDR5 4800MHz
- CPU
- 2x Intel Xeon Platinum 8480C (56C 105M Cache 2.00 GHz)
With every server
The same terms on all twenty.
Whichever chassis you pick, you get the whole node, a meter that counts seconds, root access and an invoice in euros.
billing · nvidia-dgx-b300
metering
€0.01333/s
8 × B300 · on demand · per second
0s€0.0015s€0.2030s€0.4045s€0.6060s€0.80
- 1 minute
- €0.80
- 20-min test
- €16.00
- 1 hour
- €48.00
Billing
The meter counts seconds
A full B300 server costs €48.00 an hour, charged by the second from boot to stop.
ssh · eu-ams-1
Access
Root over SSH
Your image, your drivers, your scheduler. Or start from a Nexpod Hub template.
invoice · draft
EUR
Period 01–31 Oct 2026Account · Customer name
LineHoursRateAmount
NVIDIA DGX B300131,400 s · 8× B300 · on demand36.50€48€1,752.00
NVIDIA DGX H20043,200 s · 8× H200 · on demand12.00€32€384.00
ASUS ESC N8A-E1215,300 s · 8× H100 · on demand4.25€24€102.00
Subtotal · excl. VAT€2,238.00
Invoicing
One invoice, in euros
Every server is a line: hours times its hourly rate, excluding VAT.
nvidia-smi topo -m · 8× B300
GPU0GPU1GPU2GPU3GPU4GPU5GPU6GPU7
GPU0XNV18NV18NV18NV18NV18NV18NV18
GPU1NV18XNV18NV18NV18NV18NV18NV18
GPU2NV18NV18XNV18NV18NV18NV18NV18
GPU3NV18NV18NV18XNV18NV18NV18NV18
GPU4NV18NV18NV18NV18XNV18NV18NV18
GPU5NV18NV18NV18NV18NV18XNV18NV18
GPU6NV18NV18NV18NV18NV18NV18XNV18
GPU7NV18NV18NV18NV18NV18NV18NV18X
NV18 = 18 NVLink connections between the pairGPU0 reaches all seven others directly
A whole node
Eight GPUs, no neighbours
Each listing is the full machine. All eight GPUs share one NVLink fabric and nothing is split with other tenants.
Rent a server
Found your machine? It boots in minutes.
Create an account, choose the server and region, and connect over SSH. Stop it when the job is done and billing stops with it.
On-demand ratesEUR · excl. VAT
- B300· 288 GB HBM3e€6.00/GPU·hr
- B200· 180 GB HBM3e€5.00/GPU·hr
- H200· 141 GB HBM3e€4.00/GPU·hr
- H100· 80 GB HBM3€3.00/GPU·hr