-
Products
- Compute
- Network
- Platform
- Add-ons
-
Solutions
- By Industry
- By Use Case
- Featured
-
Developers
- Docs
- Tools
- Resources
- Pricing
- Login
A100, A40 and A10 nodes
A100 80GB, A40 48GB and A10 24GB GPU based server for AI inference, fine-tuning and small teams of AI developers. Instant provision (in minutes) for zero hourly billing and availability in various jurisdictions.
A100 80GB · A40 48GB · A10 24GB · zero hourly billing
Active jobs
| Server | CPU | RAM | Storage | Network | Availability | Per month | Order |
|---|---|---|---|---|---|---|---|
| ACCELERATORAccelerator RTX6000 | AMD EPYC 7702p · 64c | 128 GB | 2x1TB NVMe | 1x 10GbE · 100 TB | 4 readyAmsterdam | €349or €314.10 billed yearly | Buy now |
| ACCELERATORAccelerator H100 NVLink | AMD EPYC 7702p · 64c | 256 GB | 2x4TB NVMe | 2x 10GbE · 100 TB | Quoted per site | Quote | Get a quote |
| ACCELERATORAccelerator A100 | AMD EPYC 7702p · 64c | 512 GB | 2x4TB NVMe | 2x 10GbE · 100 TB | Quoted per site | Quote | Get a quote |
3 plans · Monthly prices. Yearly billing takes 10% off. Need a custom build? Talk to engineering →
How your server reaches you
You order in the shop. We find or build the machine, put it in a rack and install it. Below is how that goes, using the screens from our own portal.
01
On shop.netrouting.com you pick a city first, then memory, disks, bandwidth and IP space. Prices are shown per option, per month. You only need an account when you check out.
Bandwidth is a setting on the switch port. If you start on 10G with 100 TB and later want 20G unmetered, we change the port and the server stays where it is.
02
Ready now131 ready now
These are already in a rack in that city and have been tested. Once you've paid, we assign one to you and install the OS you picked.
We take that number from the shop's inventory. It's how many are free right now.
Custom build
Custom is for configurations we do not have ready: extra memory, other disks, a RAID or ZFS layout. Our engineers build and burn in the server, which takes longer than a stock machine.
The product page says which one applies to yours.
03
In Amóni you see each server go from payment to install. The IP address is there as soon as a machine has been assigned.
No need to keep the tab open, we email you the address and root login when it's done.
04
A lot of what used to mean writing to support, you now do yourself in Amóni.
The people behind it
Our network and datacenter team is in The Hague, at DFDC THG1, which is also our European head office. Every site is run from there. When a disk has to be swapped or your BGP session set up, it's done by the people who built the network.
Network and add-ons
Anything on this list can go on an Edge server or bigger, either when you order or later on. Express machines are the exception, they come as they are. The prices here are pulled from the shop.
Edge Pro and higher get a 10G port with 100 TB outbound per month, Edge gets 1G, and inbound traffic is free on all of them. If you'd rather not count at all, take an unmetered port. That capacity is reserved for you and we don't throttle it.
At 40G it's either four 10G ports bonded with LACP or one 40G QSFP port, depending on the rack. We'll tell you which one you get.
| Port | Traffic | Per month |
|---|---|---|
| 10G | 100 TB | Included |
| 1G | Unmetered | +€100 |
| 2G | Unmetered | +€200 |
| 5G | Unmetered | +€400 |
| 10G | Unmetered | +€600 |
| 20G | Unmetered | +€1,000 |
| 40G | Unmetered, 4×10G bonded or 1×40G | +€2,000 |
For two or more servers in the same datacenter. Replication and database traffic go over the VLAN and stay off your public port.
You get one static IPv4 address with the server. Need more, add a block: anything from a /29 (5 usable) to a /26 (61 usable). Reverse DNS for those is yours to set.
A /64 for every server.
We can announce your own /24 or larger from our network. You send the LOA, we check it, then the announcement goes live.
With your own ASN you can run BGP sessions with our routers. After you open a ticket we create the IRR objects and set the sessions up, which normally takes one business day.
There's 20 Gbps of DDoS protection on every server. Premium takes that to 50 Gbps for +€50 a month.
Our engineers can manage the software for you as well, and you reach them through the same ticket queue you use for hardware.
We patch the OS and set up monitoring for you. When an alert fires we're usually already working on it before you'd have noticed anything. It's covered by the Netrouting Guarantee.
We take care of the hypervisor itself, on one host or on a cluster with Ceph distributed storage, and you keep running your own VMs on top.
We supply the cPanel license and manage the server. The license is billed on top, so ask us for a quote.
You pick add-ons on the product page when you order. After that your IP blocks and reverse DNS are in Amóni, right next to the console.
Use cases
If all your AI production workloads run fine with H100 then you only need capacity for A100 for those workloads and A40 or A10 for the rest.
The lineup
As the H100 units are for the time being somewhat of a restricted part, price tagged in a manner which does not render it very suitable for the vast majority of AI-focused production workloads, we here at Netrouting are not resellers of this GPU nor do we keep H100 units in stock for immediate deployment to customers.
Included with every plan
All GPU plans come with everything you'll need after you deploy your inference endpoint.
Why Netrouting
Common questions
Volumetric floods (e.g. UDP, ICMP and TCP floods) as well as reflection/amplification attacks (e.g. DNS, NTP, memcached etc.) and protocol based attacks (e.g. SYN floods, ACK floods etc. as well as fragmentation and malformed packet attacks). We also absorb multi-vector attacks.
In case the incoming traffic load (DDoS attack) is bigger than the traffic that can be filtered by us, then the targeted IP will be black-holed.
Dedicated DDoS Protection is automatically enabled for all relevant services such as Cloud Compute, Colocation, Bare Metal servers and GPU servers. No activation is required. The protection starts as soon as the related resource goes online.
No, Netrouting DDoS Protection is a network-layer attack mitigation service, it is designed to protect against volumetric traffic attacks (so called floods) at network-layer (so called Layer 3/4). For application-layer attacks (so called Layer 7 attacks) like HTTP floods, slowloris, etc. you would use a Web Application Firewall (WAF) or layer 7 packet processing, edge caching services like Cloudflare or Akamai.
Gaming / DDoS | Netrouting DDoS Protection - free, always-on service protecting all Netrouting infrastructure (Cloud Compute, Colocation, Bare Metal servers, GPU servers) from network-layer (L3/L4) DDoS attacks. No setup required.
Every city page now includes the public test IP address for that location, as well as a link to 100 MB, 1 GB and 10 GB test files. You can download these files from the location nearest to you with curl or wget, or download them from your browser.
Netrouting provides dedicated uplink ports (e.g. 1GbE, 10GbE). This means that there is dedicated bandwidth and a dedicated uplink port, which means that there is no loss of performance and quality caused by shared uplink usage as found with other providers.
200+ teams rely on Netrouting for bare metal, cloud, GPU servers and colocation. Easy to order servers within minutes. We also love to design custom servers for you and your business, with our experienced engineers.
We need a few cookies for the site to work at all. Those are always on. Everything else, remembering your preferences, counting how pages are used, and seeing which campaign brought you here, happens only if you allow it. You can change your mind at any time through Cookie settings at the bottom of any page. What we store, exactly
Before you go
We hold massive hardware in stock and build everything fully custom to your spec, at sharp, competitive pricing. We deliver almost anywhere, even beyond our own data centers, and we take on any challenge. Tell us what you need.