Lyceum 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
Lyceum suits teams that need current NVIDIA parts and fast interconnect without a large footprint, and it sits well for anyone billing by the second. GPU hours start at $1.19, above the €0.63 typical entry on this site, under Lyceum Technology Germany GmbH.
Key facts
- Best for
- European teams training on current NVIDIA hardware and paying by the second
- GPU models
- NVIDIA L40S, NVIDIA A100, NVIDIA H100, NVIDIA H200, NVIDIA B200, NVIDIA B300, NVIDIA GB200, NVIDIA GB300↗
- GPU hour from
- $1.19↗
- Billing granularity
- Per second↗
- Locations
- 3↗
- Interconnect
- 400 Gb/s InfiniBand NDR included on large-scale clusters↗
- Compliance certifications
- GDPR↗
- Legal entity
- Lyceum Technology Germany GmbH↗
- Headquarters
- Germany↗
- Support channels
- Email, Phone, Help-centre, Community↗
- Pricing model
- Hourly billing
- Official website
- lyceum.technology
Facts checked September 26, 2026, 5 sources
Apps and profiles
- Elsewhere
- LinkedIngithub.com
Pros and cons
Pros
- Billing is per second, a rarity among GPU clouds, so idle capacity stops costing
- The roster lists NVIDIA L40S, NVIDIA A100 and NVIDIA H100, a data centre focused set
- 400 Gb/s InfiniBand NDR included on large-scale clusters is included on large-scale clusters, which matters for distributed training
- Support covers email, phone and help centre
Cons
- Entry GPU price of $1.19 is above the €0.63 median here
- A small footprint of 3 locations offers little choice on where jobs run
- Billing offers a single hourly model, so nothing locks a rate over a long run
Who it suits
- European research groups that need current NVIDIA silicon for training runs
- Teams doing distributed training over a fast interconnect
- Anyone whose spend is bursty and who prefers paying per second to per hour
Who should look elsewhere
- Teams chasing the lowest entry rate, who may find cheaper on Salad
- Workloads needing a broad catalogue beyond data centre parts, where ComputePrices.com lists more
- Buyers who need a choice of regions, since Lyceum runs in 3
How Lyceum compares
Lyceum vs RunPod8.9 vs 9.5
Lyceum vs ScalaHosting8.9 vs 8.3Lyceum vs Contabo8.9 vs 9.2
Lyceum vs Kamatera8.9 vs 9.5
Lyceum vs Cloudways8.9 vs 9.2
Lyceum vs PlanetHoster8.9 vs 8.7
Lyceum vs Acronis8.9 vs 9.9
Lyceum vs Aruba8.9 vs 9.3
Lyceum vs Decodo8.9 vs 9.1Lyceum vs Hostinger8.9 vs 9.1
Lyceum vs MarsProxies8.9 vs 9.0Lyceum vs hide.me8.9 vs 8.7
Lyceum vs Proton Drive8.9 vs 8.3
Frequently asked questions
How much does a GPU hour cost at Lyceum?+
Entry sits at $1.19 for the cheapest card on the pricing page. Set against the €0.63 median across this comparison, that lands well above a typical low end, so shoppers on a tight budget may find cheaper elsewhere.
How is GPU time billed?+
Charges are per second, and the listed model is hourly. Per-second metering suits short tests and bursty jobs, though only 50.9% of rivals on this list bill by the hour instead, so stop-start workloads pay less.
Which GPUs are on offer?+
The roster covers NVIDIA L40S, NVIDIA A100 and NVIDIA H100. That is a data centre focused set rather than a broad marketplace, which is a narrower choice than catalogues that also carry consumer cards.
Where does Lyceum run?+
Capacity is spread across 3 locations, and the provider is Lyceum Technology Germany GmbH, headquartered in Germany. Regional choice is limited against providers running far more sites, so data placement may decide for you.
Is there a free tier or trial?+
The listed model is hourly, so spending is usage based rather than covered by an allowance. 25.3% of brands in this comparison include a free tier, which this one does not, so usage has to be budgeted from the outset.
What large-scale training hardware is included?+
Large-scale clusters come with 400 Gb/s InfiniBand NDR included on large-scale clusters. That bandwidth is the detail that matters for multi-node training, and it sits inside the offering rather than as a separate charge.
How do I reach support?+
Help runs through email, phone and help centre, so phone and email are both open alongside written material. Phone reach is not universal among competitors, with 43.6% of brands offering it.
What certifications and data rules apply?+
Lyceum states GDPR as its compliance position and lists no other certifications or licences. Teams whose procurement checklist demands attestations beyond that would need to check the documentation against their own requirements before signing anything.
Sources
- lyceum.technology/company/about checked September 26, 2026
Backs: Compliance certifications, Headquarters
“That means GDPR compliance by default, EU data sovereignty, and low-latency access for European teams.”
- lyceum.technology/company/contact checked September 26, 2026
Backs: Support channels
“Email us ... Documentation lyceum.technology/docs”
- lyceum.technology/infrastructure/large-scale-clusters checked September 26, 2026
Backs: Interconnect
“400 Gb/s InfiniBand NDR High-bandwidth, low-latency interconnects for efficient distributed training.”
- lyceum.technology/legal/imprint checked September 26, 2026
Backs: Legal entity
“Website Operator Lyceum Technology Germany GmbH”
- lyceum.technology/resources/pricing checked September 26, 2026
Backs: Billing granularity, GPU hour from, GPU models, Locations, Pricing model
“Billed per second. No minimum.”
Compiled from 5 source pages, checked September 26, 2026. We did not open an account.
Not confirmed yet: Egress fee.
Related by the facts
Lyceum
European teams training on current NVIDIA hardware and paying by the second