About 1Legion
We are a bare-metal NVIDIA GPU cloud. We started in Málaga, Spain, and are now headquartered in Vancouver, with GPU capacity in Toronto and Montreal.
1Legion rents dedicated NVIDIA GPU servers and clusters to AI companies on 36 to 60 month contracts. We build our HGX B300 clusters to a published delivery specification, burn in every cluster before handover and run it under written operating commitments.
What 1Legion does
1Legion has four lines of business: GPU servers and clusters, managed Kubernetes, colocation, and operations for companies that own GPU hardware.
GPU servers and clusters
Dedicated NVIDIA GPU servers and clusters for AI companies, rented as whole servers on 36, 48 or 60 month contracts. NVIDIA HGX B300 is our main line.
Managed Kubernetes
We install Kubernetes on your dedicated servers with the NVIDIA GPU Operator and Network Operator, then run it, upgrade it, monitor it and stay on call for it. Managed Slurm is the alternative.
Colocation
We resell data center colocation per kW on multi-year terms to companies that own their hardware.
Operations for GPU owners
We deploy and operate GPU clusters for hardware investors and for companies becoming cloud providers, and we bring them customers from our own pipeline.
Our pipeline holds more demand for GPU servers than we can deploy on our own, so we work with partners who bring capital, facilities or hardware to serve it.
Where we started and where we are now
1Legion started in Málaga, Spain, and is now a multinational company headquartered in Vancouver, Canada.
The company behind the brand is Lynx Cloud Corp., incorporated in British Columbia. Our offices are in Vancouver and San Diego.
Our servers run in data centers in Toronto and Montreal. HGX B300 capacity in Delaware, in the United States, is coming online, and a brownfield site in Finland is in development for future capacity in Europe. Our partner network adds data centers in more cities across North America, Europe, Africa and Australia. We name our data center locations by city, not by facility.
How we work
We put the specification, the operating commitments and the contract terms in writing before a customer signs.
- A published specification. Every HGX B300 proposal starts from our delivery specification, and any change you need is listed as a numbered deviation.
- Written commitments. We acknowledge an incident and assign an incident commander within 15 minutes, send updates at least every 60 minutes and deliver a written postmortem within 7 calendar days of a customer-impacting incident. The full list is in our operating commitments.
- Whole servers. We never split a server between customers. We rent 8-GPU servers, 4-GPU servers for the RTX 4000 Ada and RTX A4000, and 72-GPU racks for the GB300 NVL72.
- Long terms at one fixed rate. GPU server contracts run 36, 48 or 60 months at one fixed rate with no egress fees, and longer terms carry lower rates.
- No trials. Instead of trials or proofs of concept, every cluster passes NVIDIA's hardware validation tests HW-1 to HW-5 before handover, and you receive the results in an acceptance report. See burn-in and acceptance.
Ousia Compute, our joint venture with KIDZ AI
In May 2026, 1Legion and Classover, now KIDZ AI (Nasdaq: KIDZ), formed Ousia Compute, a dedicated GPU joint venture, and Classover selected 1Legion to deploy up to $50M of GPU infrastructure.
The joint venture was announced on May 26, 2026; read the press release on ACCESS Newswire. Our article on what the Ousia Compute joint venture means gives the background.
Who we work with
AI companies, GPU clouds, data center operators, network carriers and hardware makers we work with.
Customers and partners27 companies
Hardware partners9 server makers
Building the executive team
We are building our executive team the way the large GPU clouds did, starting with four senior roles.
- A VP of infrastructure: our top technology leader, reporting to the CEO and responsible for the infrastructure we build in Canada, the United States and Europe.
- A VP of data center development, responsible for power and sites.
- A VP of capital markets and structured finance, responsible for lender relationships, fundraising and investor presentations.
- A general counsel, to lead our legal work.
Our careers page describes each role and how to apply.
Offices
We have three offices, with our headquarters in Vancouver.
Vancouver (headquarters)
329 Howe St, Unit #688, Vancouver, BC V6C 3N2, Canada
San Diego
7950 Silverton Avenue, Suite 103, San Diego, CA 92126, United States
Email [email protected], call +1 (858) 264-1652 or use our contact page.
Common questions
Where is 1Legion based?
Our headquarters is in Vancouver, British Columbia, and we have an office in San Diego. The company started in Málaga, Spain.
Which company do customers sign contracts with?
Lynx Cloud Corp., a company incorporated in British Columbia, Canada. 1Legion is a brand of Lynx Cloud Corp.
Where do your GPU servers run?
In data centers in Toronto and Montreal. HGX B300 capacity in Delaware, in the United States, is coming online, and a site in Finland is in development. See our data centers.
Can we try the servers before we sign?
We do not offer trials or proofs of concept. Every cluster passes NVIDIA's hardware validation tests HW-1 to HW-5 before handover, and you receive the results in an acceptance report. See burn-in and acceptance.
What is Ousia Compute?
A dedicated GPU joint venture that 1Legion and Classover, now KIDZ AI, formed in May 2026. Classover selected 1Legion to deploy up to $50M of GPU infrastructure. Read the press release.
Is 1Legion hiring?
We are recruiting for four senior roles: a VP of infrastructure, a VP of data center development, a VP of capital markets and structured finance, and a general counsel. See careers at 1Legion.
Tell us what you need to run.
Share the hardware, the number of servers, the term and the start date. You receive a written proposal built on our delivery specification.