GMI Cloud, a startup that provides cloud computing infrastructure for training, deploying, and running AI models, has raised $668 million in Series B funding, founder and CEO Alex Yeh said Tuesday.
The round includes $223 million in equity funding and a $445 million credit facility, The Wall Street Journal reported. ARCHIV led the round, Yeh said, adding that Nvidia, South Korean investment firms DSC Investment and KB Investment, cybersecurity company Trend Micro, life insurer Kyobo Life Insurance and telecommunications company KT Corporation also participated.
Yeh did not disclose GMI's valuation or identify the lenders behind the credit facility. ARCHIV also does not appear to have a public website or company profile that provides further details about the firm.
"This funding lets us bring more compute online, across more of the world, for more builders," Yeh said.
What GMI Cloud does
GMI Cloud gives customers access to Nvidia GPUs and other tools needed to run AI products at scale. Customers can use the platform for large language models, image and video generation, audio models, and other AI systems. They can run their own models or open-source models.
"Dedicated inference infrastructure can provide better GPU utilization, predictable latency, and scalable deployment options compared with general-purpose cloud environments," according to GMI Cloud.
Yeh said GMI Cloud's connections with Taiwan's technology industry help it get equipment and build new computing clusters on time. He said all customer clusters have started operating as planned.
"Our customers have their own missions. Our job is to make sure compute is never the reason they fall short: secure and stable, running the most advanced models, and built by a team that knows their business," Yeh said.
GMI Cloud's contracted annual recurring revenue has reached more than nine times its level at the end of 2025, according to Yeh. Its platform now processes around 4 trillion tokens each week, he said. Yeh did not disclose specific revenue figures.
Expanding AI computing capacity
Yeh said he started GMI Cloud in 2021 to make computing capacity available to companies that could not reserve large amounts of infrastructure years in advance.
In a 2024 interview with TechCrunch, Yeh said GMI Cloud first provided computing services for bitcoin before moving into AI cloud infrastructure as demand for GPUs grew.
GMI Cloud previously raised $82 million in Series A funding in October 2024. The round included $15 million in equity funding and $67 million in debt financing, taking its total capital at the time to more than $93 million.
GMI Cloud operates data centers in the United States and several Asian markets. Last November, GMI Cloud announced a $500 million AI data center project in Taiwan using Nvidia's Blackwell GB300 chips. The company said the facility would have about 7,000 GPUs and use around 16 megawatts of power.
GMI Cloud has also announced plans to build more AI infrastructure in the U.S. and Japan.