GMI Cloud Raises $668M With NVIDIA Backing to Scale GPU Capacity Across US, Taiwan, and APAC
GMI Cloud announced $668 million in new financing on October 1, the largest capital raise in the company’s history. The round comprises $223 million in equity through a Series B and a $445 million credit facility led by CTBC.
The Series B was led by ARCHIV, a new San Francisco-based investment firm specialising in AI and robotics. NVIDIA joined as a participant alongside a cluster of Asia-Pacific institutional investors including DSC Investment, KB Investment, Kyobo Life, KT Corporation, and Trend Micro.
Traction Numbers
GMI Cloud’s contracted ARR has reached more than 9x its year-end 2025 level. Live ARR in production has grown more than 4.5x over the same period. The company’s inference platform processes approximately 4 trillion tokens per week, a figure that reflects its position as a provider to inference-layer customers rather than just raw compute renters.
Notable customers include Fireworks, Nous Research, OpenRouter, Cartesia, Higgsfield, Reflection, and Utopai Studios. Several of those are themselves infrastructure-layer companies, meaning GMI Cloud’s capacity underpins multiple rungs of the AI stack.
Expansion Plans
Funding goes toward GPU capacity expansion in the United States, Taiwan, and the rest of APAC. The company already operates a Taiwan AI Factory, announced in 2025 as its first facility in Asia. A Japan sovereign AI initiative was announced earlier in 2026.
CEO Alex Yeh framed the Taiwan position as a supply chain advantage rather than just a manufacturing location: “In AI infrastructure, a delivery date is a promise. Customers plan launches, hiring, and revenue around it. Our place in Taiwan’s supply chain is how we keep that promise, cluster after cluster globally.”
Infrastructure Round Context
The $445 million credit facility follows a pattern seen across the GPU infrastructure sector in 2026, where equity rounds are supplemented by debt to fund capital expenditures without diluting equity holders. CTBC’s participation signals continued appetite from Asian financial institutions to finance AI infrastructure at scale.
NVIDIA’s equity participation in a GPU cloud customer reinforces a pattern of the chipmaker taking direct stakes in operators that deploy its hardware at high volume, creating alignment around product roadmap and allocation priorities.
GMI Cloud’s per-week token count, at 4 trillion, is a comparable public figure to other inference-layer operators. It converts to roughly 570 billion tokens per day across its customer base.