GMI Cloud
Category: AI Infrastructure
AI-native neocloud providing NVIDIA GPU clusters and production AI inference services across the US and Asia-Pacific. GMI Cloud was founded in 2021. The company is led by Alex Yeh. Based in Mountain View, United States. Total funding raised: $750M. Latest round: Series B. Key investors include ARCHIV, NVIDIA, Headline Asia, Banpu Next, Wistron, CTBC Bank, DSC Investment, Trend Micro, KB Investment, Kyobo Life, KT Corporation.
- Founded
- 2021
- Headquarters
- Mountain View, United States
- Total funding
- $750M
Value proposition
A vertically integrated AI infrastructure stack covering hardware, compute, orchestration and inference. It delivers predictable performance, scalable capacity and cost-efficient production AI inference on owned NVIDIA GPU data centers.
Products and solutions
Dedicated and on-demand NVIDIA GPU clusters (H100, H200, Blackwell GB200/GB300 NVL72), Inference Engine and serverless inference APIs, GMI Studio, Kubernetes-based Cluster Engine orchestration, AI agent infrastructure, Taiwan AI Factory, Japan sovereign AI initiative.
Unique value
One of only about six NVIDIA Reference Platform Cloud Partners globally. A single platform spans US and APAC regions, so customers can place workloads near their users and data. Close ties to Taiwan's AI server supply chain shorten time from order to deployment.
Target customer
AI-native companies, model providers, hyperscalers, AI developers and enterprise AI teams that need dedicated or on-demand GPU capacity in the US and APAC.
Industries served
AI model developers, generative media, inference platforms, cybersecurity, enterprise AI, sovereign AI in APAC
Technology advantage
Full-stack ownership from hardware to inference layer, 30,000+ GPUs deployed and 99.99% availability. Supply-chain proximity to Taiwanese server makers (Wistron) gives predictable delivery timelines. The inference platform processes about 4 trillion tokens per week.
How they differentiate
It focuses on production inference rather than only training clusters. It offers dual US and APAC footprints (Taiwan, Japan) for data-residency and sovereign AI needs, and NVIDIA reference-architecture status. Taiwan supply-chain links help secure GB200/GB300 capacity early.
Main competitors
CoreWeave, Lambda, Nebius, Together AI
Key partnerships
NVIDIA (Reference Platform Cloud Partner and investor), Wistron, Banpu Next, CTBC Bank (credit facility), VAST Data (storage), Google Cloud, Trend Micro
Notable customers
Fireworks AI, Higgsfield, Nous Research, OpenRouter, Reflection AI, Cartesia, Trend Micro, Utopai Studios
Major milestones
2021: founded, initially as crypto mining infrastructure, then pivoted to an AI GPU cloud. Oct 29, 2024: announced $82M financing, comprising $15M Series A equity led by Headline Asia and $67M debt; reported cumulative capital exceeding $93M. 2025: announced Taiwan AI Factory, its inaugural facility in Asia. 2026: Japan sovereign AI initiative. Sep 30, 2026: announced $668M financing ($223M Series B led by ARCHIV with NVIDIA, plus a $445M CTBC-led credit facility). The 2024 and 2026 financing packages total $750M, including $512M of debt financing; $750M is not a verified lifetime funding total.
Growth metrics
Contracted ARR is above $600M, more than 9x the end-2025 level. Production ARR grew more than 4.5x over the same period. The platform processes about 4 trillion tokens per week, runs 30,000+ GPUs and serves 300+ AI team customers. Figures are as of Sep 2026.
Market positioning
A fast-growing Nvidia-backed neocloud positioned as the US–Asia bridge for AI compute and inference.
Geographic focus
United States, Taiwan, Japan, Asia-Pacific, Europe
About Alex Yeh
Ex-Mesh Ventures Analyst; Director at Globaltec Capital; Johns Hopkins University. Founded GMI Cloud in 2021, pivoting it from Bitcoin mining infrastructure to an AI GPU cloud.
Latest news about GMI Cloud
More AI Infrastructure companies
Official website: https://www.gmicloud.ai