Kentino AI Node H100 — 8× NVIDIA H100 SXM5, 640 GB HBM3
Enterprise-grade platform, assembled and tested in the EU.
Kentino AI Node H100
640 GB HBM3, NVSwitch-Pooled
8x H100 SXM5 | 15,832 INT8 TOPS
The Hopper reference node - the best-understood 8-GPU platform in production.
An 8U node carrying eight NVIDIA H100 SXM5 80 GB modules on a single NVLink/NVSwitch baseboard: 640 GB of HBM3 at 3.35 TB/s per GPU, full any-to-any 900 GB/s GPU-to-GPU bandwidth, dual Intel Xeon Gold 6538Y+, 2 TB DDR5 ECC RDIMM and U.2 NVMe storage. The H100 is the most widely benchmarked 8-GPU platform in existence, which matters more than it sounds: if you have an existing training recipe, a vendor SLA or a procurement process built around H100 numbers, this is the node that matches them without re-validation.
Hardware
| Component | Detail |
|---|---|
| GPUs | 8x NVIDIA H100 SXM5 80 GB HBM3 (Hopper, 700 W per module) |
| VRAM pool | 640 GB HBM3, NVSwitch-pooled |
| Memory bandwidth | 3.35 TB/s per GPU |
| GPU interconnect | NVLink 4 via NVSwitch, 900 GB/s any-to-any |
| Compute | 7,912 TFLOPS bf16 dense | 15,832 TFLOPS fp8 dense | 15,832 INT8 TOPS |
| CPUs | 2x Intel Xeon Gold 6538Y+, 32C/64T each (64C/128T). Platinum 8558 (96C/192T) available as an upgrade |
| System memory | 2 TB DDR5 ECC RDIMM (16x 128 GB) |
| Storage | 2x 1.92 TB U.2 NVMe boot + 2x 3.84 TB U.2 NVMe data |
| Networking | 2x NVIDIA ConnectX-7. Up to 8x on request |
| Chassis | 8U rack-mount, front-to-back airflow, 24/7 rated |
| Power | 6x 3000 W (3+3) redundant Titanium |
What it runs
640 GB covers the whole current open-weight landscape at fp8 or Q4. Estimates are for this 8-GPU configuration with tensor parallelism.
| Component | Detail |
|---|---|
| Llama 3.3 70B / Qwen3-72B | fp8. ~90-130 tok/s single stream, 5,000+ tok/s batched |
| DeepSeek-V3.2 / R1 671B | Q4 comfortably; fp8 needs weight streaming at this pool size |
| Qwen3-235B-A22B MoE | fp8 native with context headroom |
| Qwen3-Coder-480B-A35B | Q4 agentic coding, full 8-GPU tensor parallel |
| Full fine-tune 70B | 8-way tensor parallel |
| Mixed multi-tenant | several 70B-class models resident simultaneously |
Throughput figures are estimates derived from published vendor specifications and community benchmarks for this GPU generation, not measurements taken on this exact configuration. We will give you measured numbers for your model and context length before you commit.
Power and cooling
Roughly 6.7 kW sustained - identical envelope to the H200 node. Dedicated 32 A three-phase feed per node, rack budget around one node per 10 kW, contained cold aisle below 24 C with genuine front-to-back airflow for 24/7 operation.
Export control - read this before ordering
The H100 is an Advanced Computing Item (ECCN 3A090.a). Shipment requires a completed KYC form and an End User Certificate naming the end user, the installation address and the intended use.
Budget several weeks for clearance on top of hardware lead time. If the end user cannot be named, this hardware cannot be supplied.
What you get from Kentino
Kentino supplies, configures and supports this node as a complete deployment: rack and power design against your site's actual feed, front-to-back airflow sizing for 24/7 operation, OS and driver stack (Ubuntu with pinned NVIDIA driver, CUDA, container toolkit), the inference or training stack of your choice, burn-in with a written report, EU-based logistics and warranty.
We also handle the export paperwork with you rather than mailing you a form and hoping. That part takes longer than the hardware and is where most first-time buyers of this class of machine get stuck.
Price shown is EUR excluding VAT, EXW Prague (Incoterms 2020). Collection, freight, insurance and any destination import duties are for the buyer’s account. Lead time on request.
Before you order this, ask us to quote the H200 node alongside it - at current channel pricing the 8x H200 baseboard often lands at or below the H100 board while carrying 1,128 GB instead of 640 GB. We would rather tell you that up front than sell you the wrong node.
The questions buyers ask us most often before ordering a server.
How long does it take?
Machines built from components we hold ship quickly; anything requiring a specific GPU generation depends on supply. We give you a date before you pay, and if it moves we tell you rather than letting you find out.
Can the configuration be changed before you build it?
Almost always. GPUs, memory, storage and cooling are chosen per order, and the listed configuration is a starting point rather than a fixed package. If you need more VRAM, faster storage or a different cooling approach, say so before you order and we will quote the change.
Can I collect the server in person?
You can. Our warehouse is in Prague, and collection in person is welcome — most people who come use the visit to go through the machine with our engineer and ask the questions that are awkward over email. For orders within the Czech Republic we also try to deliver personally and walk you through the setup on site.
Can I talk to someone who actually understands the workload?
Yes. We have an engineer who works on AI systems specifically, not a general sales desk. If your question is about batch sizes, quantisation, interconnect or where your bottleneck will be, ask it — that conversation usually changes the configuration for the better.
Which model can I run on this configuration?
Yes. Every machine is assembled, burn-in tested and benchmarked on real AI workloads before it ships, and it leaves us with an LLM already installed and running. You plug it in, connect it to your network and start work — the only decision left is which project it runs first.
How do you test a server before shipping?
We run it against actual AI workloads rather than synthetic scores: inference throughput, sustained load behaviour and thermals under continuous operation. You get the benchmark results with the machine, so the performance you were promised is the performance you can verify on day one.
Is the server ready to run when it arrives?
Yes. Every machine is assembled, burn-in tested and benchmarked on real AI workloads before it ships, and it leaves us with an LLM already installed and running. You plug it in, connect it to your network and start work — the only decision left is which project it runs first.
Ships from our EU warehouse. Heavy items may require freight arrangement — contact us for a shipping quote and lead time. 2-year limited warranty with advanced RMA support; extended warranty available.
Not exactly what you need?
Tell us your workload and we'll spec this platform around it — GPUs, memory, storage and cooling matched to what you actually run.