SupermicroSYS-C542i-11302U

A tower with an Intel Core Ultra 7 265, 32 GB of DDR5 and a 1,000 W power supply, delivered without a graphics card so you can choose an RTX 5070, 5080 or 5090 to suit the models you want to try.

Ask for price and lead timeSee the configuration
Supermicro SYS-C542i-11302UTower · Gold Series G2

The exact configuration

Official datasheet

Front view diagram: hover over each component.
Processor1× Intel Core Ultra 7 26520 cores · 5.3 GHz
GPUReady for NVIDIA RTX 5070 / 5080 / 5090 (optional)
Memory32 GBDDR5-5600
Storage1 TB NVMe
Network1 GbE, Thunderbolt 4 / USB4, front USB-C 20G, Wi-Fi 7, Bluetooth 5.4
Form factorTower
Warranty
1 year parts and labour
On-site engineer
Optional: extend the warranty to 3 years with an engineer next business day

A desk for developing and testing AI locally

Supermicro pitches it as the AI PC for SMEs and developers: the desk where someone tries out a model, builds an agent workflow or automates a process before moving it to a server, without sending company data to an outside service. With an RTX 5090 and its 32 GB of memory, mid-sized quantised models fit, the kind used to test an idea, and the 1,000 W supply covers that card without needing a swap.

For production it falls short. The memory is non-ECC DDR5, up to 256 GB across four slots, and the warranty is one year, so if the model is going to serve a whole department it should move to a rack machine such as the SYS-212GB-FNR-01-G2. And if you will be working with models larger than 32 GB, no card in this range will be enough.

What we do

We set it up with Ubuntu, the NVIDIA drivers and CUDA, and the same vLLM containers you would use on the server, so whatever works on the desk moves to production without being rebuilt. If it is for a small team, we make it reachable over the office network so several people can run tests against it.

Compared with the rest of the family

ModelWhat it hasWhen to choose it
SYS-822GS-NB3RT-01-G2HGX B300 8-GPU · 2 TBThe HGX B300 with Intel; the choice if your standard is Xeon.
AS-8126GS-NB3RT-01-G2HGX B300 8-GPU · 3 TBSame HGX B300 with AMD, 3 TB and a separate storage network.
AS-5126GS-TNRT-01-G22x RTX PRO 6000 · 1.5 TBTwo GPUs today, space for eight and memory to spare.
SYS-212GB-FNR-01-G22x RTX PRO 6000 · 512 GBThe most compact: 2U for a two-GPU service.
SYS-422GA-NRT-01-G24x RTX PRO 6000 · 1 TBFour GPUs that MIG splits into up to sixteen instances for several teams.
SYS-422GA-NRT-02-G28x RTX PRO 6000 · 1.5 TBThe most PCIe can do: eight GPUs and 400 GbE networking.
ARS-511GD-NB-LCC-01-G2Grace + B300 · 748 GBA GB300 node in a tower for teams that fine-tune models without booking a cluster.

See all AI and GPU servers

What the delivery includes

From the factory

  • Configuration assembled and tested by Supermicro
  • Warranty: 1 year parts and labour

From us

  • Operating system, drivers and AI stack installed
  • Testing with your workload before delivery

If you need it

  • 24×7 on-call support with SLA
  • Warranty extended to 3 years with an on-site engineer next business day

How we build it

What runs on top of these servers, explained in detail:

All our AI services

Support and training for your team

Independent support from the people who installed itLocal LLM and GPU infrastructure courseDeep learning course with PyTorch and TensorFlowRed Hat OpenShift training

Questions about the SYS-C542i-11302U

Does it come with a graphics card?

No. The Gold configuration ships without a GPU and is ready for an NVIDIA RTX 5070, 5080 or 5090, which we add depending on what you plan to do. For AI we usually recommend the 5090 for its 32 GB of memory, double the 16 GB of the 5080, because the size of model you can load depends mostly on that figure.

Which AI models can I run on an RTX 5090?

Models of up to around 30 billion parameters quantised to 4 bits, or somewhat smaller at 8 bits, depending on how much context you need, which is enough for prototypes, coding assistants and agent testing. A 70-billion-parameter model no longer fits in 32 GB; for that you need the 96 GB RTX PRO 6000 cards in the rack servers.

Can I leave a model running in production on it?

For a pilot with a few people, yes; for production we do not recommend it. The memory is non-ECC DDR5 (ECC being the kind that corrects bit errors), the warranty is one year and it has a single 1,000 W supply, so once the model starts serving a department we move it to a rack machine such as the SYS-212GB-FNR-01-G2 with the same containers.

Is the SYS-C542i-11302U right for what you're building?

Tell us which models you'll run and who will be working on it, and we'll come back with the price, a confirmed lead time and what it would take to get it running.

Ask for price and lead timeWhatsApp+34 91 198 02 43