GMI Cloud Raises $223M Series B Led by ARCHIV to Scale NVIDIA-Standard AI Cloud

GMI Cloud Secures $223M Series B

GMI Cloud, an AI infrastructure-as-a-service company, has raised a $223 million Series B round led by ARCHIV. The funding was announced on September 30, 2026.

What GMI Cloud Does

GMI Cloud operates an on-demand GPU cloud platform purpose-built for AI training and inference workloads. The platform is built on NVIDIA H100, H200 and Vera Rubin systems and is orchestrated by the company's Kubernetes-based Cluster Engine.

The company operates dozens of data centers across Taiwan, Thailand and Malaysia. It is expanding its United States presence and runs a sovereign AI initiative in Japan. GMI Cloud is one of a small group of cloud providers built entirely to NVIDIA standards.

Founded in 2021 and led by CEO Alex Yeh, GMI Cloud serves enterprise customers including Higgsfield, OpenRouter, Nous Research, Fireworks AI and Trend Micro.

The Round

• Round type: Series B
• Amount: $223,000,000 USD
• Announced: September 30, 2026
• Lead investor: ARCHIV
• Participating investors: Kyobo Life Insurance Group, NVIDIA, KB Investment, DSC Investment, ARCHIV, KT Corporation, Trend Micro

Why This Matters for the AI Infrastructure Market

The Series B places GMI Cloud among the better-capitalized independent GPU cloud providers at a time when demand for AI training and inference capacity continues to outpace supply. The investor list combines strategic and financial backers across the United States and Asia, including NVIDIA, KT Corporation and Trend Micro.

GMI Cloud's positioning is distinct in two ways. First, its stack is built entirely to NVIDIA standards, spanning current-generation H100 and H200 systems as well as the Vera Rubin platform. Second, its geographic footprint is concentrated in Asia — Taiwan, Thailand and Malaysia — with an active expansion into the United States and a sovereign AI initiative in Japan.

Sovereign AI has become a priority for governments and enterprises seeking to keep data and compute within national borders. GMI Cloud's Japan initiative targets that demand directly, while its existing data center network provides the physical capacity to serve regional enterprise workloads.

Enterprise Traction

The company's customer roster spans generative media, model routing, open research and cybersecurity. Named customers include Higgsfield, OpenRouter, Nous Research, Fireworks AI and Trend Micro. These are workloads that typically require sustained GPU capacity for both training and inference, which aligns with GMI Cloud's on-demand model and its Kubernetes-based orchestration layer.

What to Watch

With $223 million in new capital, GMI Cloud is positioned to accelerate data center expansion, deepen its United States presence and continue building out sovereign AI capacity in Japan. The involvement of NVIDIA as an investor is notable given the company's NVIDIA-standard infrastructure, and the participation of KT Corporation and Trend Micro signals continued strategic interest from Asian telecom and cybersecurity players.

For enterprises evaluating GPU cloud providers, GMI Cloud now has a larger balance sheet behind its capacity commitments. For the broader AI infrastructure market, the round is another signal that capital continues to flow toward providers that can deliver NVIDIA-standard compute at scale across multiple regions.