Lenovo ThinkEdge servers
For processing video and sensor data, or running AI models, where the data is created, in the factory, the shop or the base station, when latency matters or the link to the data centre isn't reliable. The SE455 V3 fits NVIDIA GPUs into a shallow cabinet and copes with dust, heat and vibration.
ThinkEdge SE455 V3Choose where to process
At the site
A 2U server 438 mm deep with up to six L4s or two L40S GPUs, for inference next to the cameras or the machines. It runs from 5 to 55 °C depending on the configuration.
In your data centre, if the link allows it
If the network between your sites and the data centre can carry the volume, centralising inference on a four- or eight-GPU server is easier to maintain.
Before you ask for a price
At the site or in the data centre?
Process at the site if you need a response in milliseconds, if the link drops or if sending all the video to the data centre costs more than the server. Otherwise, centralise: one GPU server is easier to look after than ten spread across sites.
What your data centre needs
The SE455 V3 is 438 mm deep and fits two-post or 600 mm racks. It takes two 1,100 or 1,800 W power supplies, or -48 V DC ones for telecom sites. With L40S GPUs the room must stay below 40 °C; with L4s it copes with up to 55 °C in limited configurations.
Comparison
| Model | GPU | Processor | Max. memory | Form factor | Cooling |
|---|---|---|---|---|---|
| ThinkEdge SE455 V3 | Up to 6 NVIDIA L4 or 2 L40S | 1 AMD EPYC 8004, up to 64 cores | 768 GB | 2U, 438 mm deep | Air, with dust filter |
| ThinkSystem SR650a V4 | Up to 4 RTX PRO 6000 (96 GB) at 450 W, or 2 at 600 W; also H200 NVL and L40S | 1 or 2 Intel Xeon 6, up to 86 cores | 8 TB | 2U | Air |
| ThinkSystem SR675 V3 | Up to 8 RTX PRO 6000, H200 NVL or L40S | 1 or 2 AMD EPYC 9005 or 9004, up to 160 cores | 6 TB | 3U | Air |
Why buy it from us
The server arrives at each site installed and tested with your application, and if it ever fails, you call the same team that put it into service.
We start from what will run at each site: how many cameras or sensors, which AI model, what latency you need and where the server will sit. That tells us whether two L4s are enough or you need more GPUs, and whether to process at the site or in your data centre.
Before it ships, we configure it with the operating system, Kubernetes or the hypervisor and your application, and send it ready to plug in. At the site, all it needs is racking and a network connection.
When something breaks you talk to the engineers who commissioned that machine, and if you need it there's 24×7 on-call cover with a contractual SLA. It's the same independent support we provide for IBM Power, Linux and Ceph estates.
We're an IBM Business Partner and partners of Red Hat, SUSE, Canonical and CIQ. With CIQ we install Rocky Linux for AI, with the NVIDIA driver and CUDA already validated, on the same servers.
We work in all three languages, support included, so if your company has teams in several countries they deal with the same group of engineers.
How we build it
What runs on top of these servers, explained in detail:
Support and training for your team
Frequently asked questions
What is an edge server?
One that sits where the data is created, in a factory, a shop or a base station, rather than in the data centre. It pays off when latency matters, when the link to head office isn't reliable or when there's too much data to upload, as with video from many cameras.
Will it cope in a warehouse without air conditioning?
In limited configurations it runs from 5 to 55 °C, has a dust filter with an airflow sensor and is tested against shock and vibration. From 50 °C upwards it only takes A2 or L4 GPUs. We help you choose the configuration for the site.
What happens if it's stolen or opened?
The SE455 V3 detects when the cover is opened and when the unit is moved, and can shut down, lock the encrypted drives and cut off remote access. We configure it before shipping.
Which GPUs does it support?
Up to six single-slot NVIDIA L4 or A2 GPUs, or two double-width L40S or L40 GPUs. Lenovo's guide doesn't include the RTX PRO Blackwell cards: if you need them, the model is the SR650a V4, which goes in a data centre.
Do you have list prices?
No, because the price depends on the processor, GPUs, memory, network and warranty. We send you a quote with a confirmed lead time for the configuration that fits, and if you have several sites, for all of them.
What does each site need?
Tell us how many sites you have and what has to run at each one, and we'll come back with the configuration, the price and a confirmed lead time.
Blog!
Our mission
SIXE ensures that your infrastructures based on IBM®, Red Hat®, Canonical®, and SUSE® technologies remain stable, secure, and optimized for as long as you need them.
At SIXE, we are committed to sustainability. That’s why we help you design, implement, and maintain durable and efficient technological solutions.
