GMO Internet Group review
Updated September 28, 2026 · DeployCue Editorial Team · Facts checked September 26, 2026
We earn a commission when you sign up through links on this site. Partner status can affect where a brand appears in our rankings. Partners with a live offer in your country come first, then every other brand by score.
Our verdict
GMO Internet Group suits developers and small teams who want a low entry price and no transfer line item. The VPS plan starts at ¥460 with 512 MB RAM, 1 vCPU, 30 GB SSD, and the operator behind it, GMO Internet, Inc., has been running since 1991.
Key facts
- Best for
- A cheap VPS with unlimited transfer
- GPU models
- NVIDIA HGX B300, NVIDIA HGX H200↗
- Egress fee
- free↗
- Locations
- 1↗
- Interconnect
- NVLink, NVIDIA Spectrum-X↗
- Legal entity
- GMO Internet, Inc.↗
- Headquarters
- Japan↗
- Founded
- 1991↗
- Support channels
- Email, Help-centre↗
- Pricing model
- Hourly billing
- Official website
- internet.gmo
VPS
- Entry price
- ¥460↗
- Entry plan
- 512 MB RAM, 1 vCPU, 30 GB SSD↗
- Bandwidth allowance
- Unlimited data transfer (free)↗
- Billing models
- Hourly, Monthly, Reserved↗
- Server types
- VPS↗
- Operating systems
- Linux, Windows↗
Facts checked September 26, 2026, 6 sources
Apps and profiles
- Elsewhere
- Wikidata
Pros and cons
Pros
- VPS entry plan at ¥460 with 512 MB RAM, 1 vCPU, 30 GB SSD
- Unlimited data transfer (free) on VPS plans, plus free egress on the GPU service
- Billing runs hourly, monthly and reserved, so short tests need no long commitment
- GPU hosts use NVLink, NVIDIA Spectrum-X for inter-GPU and network links
Cons
- GPU hardware is limited to NVIDIA HGX B300 and NVIDIA HGX H200.
- Everything runs from 1 location, with no second region to fail over to.
- Support runs on email and help centre only.
Who it suits
- Small teams and developers who want a cheap VPS with Unlimited data transfer (free)
- Users in and around Japan who prefer a local operator, since the company is based in Japan
- GPU projects that only need recent high-end models, NVIDIA HGX B300 and NVIDIA HGX H200
Who should look elsewhere
- Teams that need a longer hardware list: SberCloud covers B300, H100 and A100.
- Multi-region deployments: ComputePrices.com tracks providers across many more sites, while this one runs 1 location.
- Experiments on consumer and desktop cards: Salad lists NVIDIA RTX 3060, NVIDIA RTX 3070 and NVIDIA RTX 3070 Ti.
What it does well
The GPU service offers NVIDIA HGX B300 and NVIDIA HGX H200 hardware, and NVLink, NVIDIA Spectrum-X handles the links between those GPUs and the network fabric. Everything runs from 1 location, which keeps latency predictable for Japan-based workloads and makes capacity planning simpler. The trade-off is a short hardware list rather than a deep catalogue of accelerators.
Who runs it
The service is run by GMO Internet, Inc., headquartered in Japan and founded in 1991. That makes it a long-established Japanese operator rather than a new entrant, and the same company fronts both the GPU and the VPS line. Support runs on email and help centre, enough for self-serve work but thin if you need voice help.
How GMO Internet Group compares
GMO Internet Group vs RunPod8.9 vs 9.5
GMO Internet Group vs ScalaHosting8.9 vs 8.3
GMO Internet Group vs Contabo8.9 vs 9.2
GMO Internet Group vs Kamatera8.9 vs 9.5
GMO Internet Group vs Cloudways8.9 vs 9.2
GMO Internet Group vs PlanetHoster8.9 vs 8.7
GMO Internet Group vs Acronis8.9 vs 9.9
GMO Internet Group vs Aruba8.9 vs 9.3
GMO Internet Group vs Hostinger8.9 vs 9.1
GMO Internet Group vs Decodo8.9 vs 9.1
GMO Internet Group vs MarsProxies8.9 vs 9.0
GMO Internet Group vs hide.me8.9 vs 8.7
GMO Internet Group vs Proton Drive8.9 vs 8.3
Frequently asked questions
What is included in the cheapest GMO Internet Group VPS plan?+
The entry plan is 512 MB RAM, 1 vCPU, 30 GB SSD and it costs ¥460. That footprint suits staging boxes, side projects and light sites rather than busy production servers. Data transfer on the VPS side is Unlimited data transfer (free), so traffic volume is not what drives the bill.
How does billing work at GMO Internet Group?+
Billing runs hourly, monthly and reserved, so you can start with the smallest commitment for a test and lengthen it once the workload is steady. The VPS pricing model is hourly, which suits short experiments, and a reserved term locks in capacity for a plan you know you will keep.
Is there an egress or data transfer charge?+
Egress is free on the GPU service, and VPS plans carry Unlimited data transfer (free). Both sides of the offering therefore keep transfer out of the bill, which matters for jobs that pull datasets out of the cluster and hand finished results back to the internet.
Which GPUs are available?+
The GPU range is NVIDIA HGX B300 and NVIDIA HGX H200. These are current high-end parts, which suits training and inference on recent models. The list is short, so confirm it covers your workload before you build a plan around the service.
Where do the servers run?+
Everything runs from 1 location, and the company behind it is based in Japan. That is convenient for users in and around Japan, with short transfers and predictable routing. For teams elsewhere it means there is no second region to spread data across.
How do you get help?+
Support runs on email and help centre. Email and a help centre handle configuration questions and billing, which covers most self-serve work. Phone is not among the listed channels, so teams that lean on live voice support should weigh that first.
Which operating systems can I run?+
The VPS plans offer linux and windows, so both Linux and Windows workloads fit. The server types here are vps, which makes the VPS line the general-purpose side, while the GPU service handles accelerated jobs on its own hardware.
Sources
- gpucloud.gmo checked September 26, 2026
Backs: GPU models
“NVIDIA HGX B300 / H200”
- gpucloud.gmo/faq checked September 26, 2026
Backs: Egress fee, Locations, Support channels
“Data transfer fee is free. No cap or surcharge on outbound transfer”
- gpucloud.gmo/services/bare-metal checked September 26, 2026
Backs: Interconnect
“8 x NVIDIA GPU per node connected internally via NVLink”
- internet.gmo/company/outline checked September 26, 2026
Backs: Headquarters, Legal entity
“Cerulean Tower, Shibuya-ku, Tokyo”
- businesswire.com/news/home/20250109879608/en/GMO-Internet... checked September 26, 2026
Backs: Founded
“Founded in 1991”
- conoha.jp/en/vps checked September 26, 2026
Backs: Bandwidth allowance, Billing models, Entry plan, Entry price, Pricing model, Operating systems, Server types
“Windows Server / LAMPサーバー”
Compiled from 6 source pages, checked September 26, 2026. We did not open an account.
Not confirmed yet: GPU hour from, Billing granularity.