The Pulse
GMI Cloud Raises $668 Million for U.S. and Asia Expansion
GMI Cloud announced $668 million in financing, combining a $223 million Series B led by ARCHIV with a $445 million credit facility led by CTBC. The company says it will use the funding to expand GPU capacity in the United States and Asia as
AI.info Team · Written with AI ·
GMI Cloud is promising customers that it can deliver computing capacity on schedule, while a customer says late-arriving capacity is unusable. The Mountain View company announced $668 million in new financing on September 30 to expand GPU infrastructure across the United States and Asia. The package combines equity with a large credit facility, as GMI Cloud reports rising contracted revenue and growing inference traffic.
A $668 Million Package, Split Between Equity and Credit
The financing includes $223 million in Series B equity and a $445 million credit facility led by CTBC. ARCHIV, a San Francisco investment firm focused on AI and robotics, led the equity round, with NVIDIA participating. Other named investors include DSC Investment, Trend Micro, KB Investment, Kyobo Life and KT Corporation; GMI Cloud’s announcement also refers to additional participants.
The company says the capital will support GPU capacity expansion in the United States, Taiwan and the broader Asia-Pacific region. It also plans to grow its inference services and hire strategically. The release does not disclose a company valuation or specify how much of the credit facility GMI Cloud has drawn.
Expansion Across Taiwan, Japan and the U.S.
The new plans build on GMI Cloud’s Taiwan AI Factory, which the company says it announced in 2025 as its first facility in Asia, and a Japan sovereign AI initiative announced earlier in 2026. The company describes Taiwan’s concentration of AI-server manufacturing as a supply-chain advantage, giving it closer relationships with manufacturers and a more predictable path from orders to deployment.
That delivery claim speaks to a practical concern for customers committing to AI deployments: infrastructure must arrive when planned. “In AI infrastructure, a delivery date is a promise. Customers plan launches, hiring, and revenue around it,” said Alex Yeh, GMI Cloud’s founder and CEO. He said the company’s Taiwan supply-chain relationships help it deliver clusters across markets.
Fireworks’ Customer Test Meets GMI’s Delivery Promise
Chenyu Zhao, co-founder of Fireworks, offered a customer’s view of the stakes: “Capacity that arrives late is capacity we can't use.” Zhao said GMI Cloud had been one of Fireworks’ strongest and most reliable providers for NVIDIA GB200 and GB300 NVL72 systems. The endorsement is specific to Fireworks’ experience; the announcement provides no independent comparison of GMI Cloud’s delivery performance.
GMI Cloud says its contracted annual recurring revenue has grown to more than nine times its year-end 2025 level, while live revenue from services in production has increased more than 4.5 times over the same period. The company says its inference platform processes about 4 trillion tokens per week. It names Fireworks, Higgsfield, Nous Research, OpenRouter, Reflection, Cartesia, Trend Micro and Utopai Studios among its customers.
More Compute Depends on Getting Capacity Online
The financing gives GMI Cloud a mix of equity and credit to pursue a buildout on both sides of the Pacific. Its stated goal is to serve U.S. companies seeking capacity in the United States and Asia, while offering Asia-Pacific enterprises infrastructure close to their users and data. Whether the expansion meets the delivery schedules customers need will matter as much as the amount of financing: the customer quoted in the announcement says capacity that arrives late cannot be used.