Dedicated GPU instances deployed in under 2 minutes. No hidden fees, no shared tenancy. Priced transparently — targeting 30% below the competition.
Trusted by researchers at
























Prices benchmarked against the market · We target 30% below market rate
Blackwell · 96 GB VRAM
Reference range: NT$27~42/hr
Volta · 128 GB VRAM
Reference range: NT$32~52/hr
Need a specific GPU, multi-node cluster, or long-term contract? Contact us to discuss your requirements.
From GPU selection to live SSH — three steps, under two minutes
Browse the catalog and choose the model and specs that fit your workload. Availability and pricing update in real time.
Provisioning starts immediately after you submit. SSH access is ready before your coffee brews — no queues, no waiting lists.
Billed hourly, charged per second. Terminate the moment your job finishes — no minimum fees, no penalties.
Our team comes from an academic research background — we know what researchers actually need. From official tax invoices to academic discounts, GPUtw.ai fills the gap that international cloud providers leave behind.
Data stays in Taiwan. Low latency, fully compliant.
Mandarin support, Taiwan-timezone service.
Issue triplicate uniform invoices for schools and companies.
Academic backgrounds on our team — we understand research workflows.
Institutions and research groups can negotiate custom rates.
Convert a personal account into a team account, invite collaborators, and let member deployments bill to one managed balance with owner-controlled limits.
Member usage can draw from the team owner balance for cleaner accounting.
Set monthly spend limits, instance caps, and deploy suspension per member.
Choose whether members can see the available team balance.
Sign up free. Pay only for what you use — billed by the second, no minimum fees.
Create Free Account