Add preferred source
KB Investment, DSC Investment, Kyobo Life Insurance, and KT have joined Nvidia in a Series B funding round for GMI Cloud, a U.S.-based AI-native cloud company. The total round size is $668 million (approximately ₩910 billion), comprising $223 million in equity investment and a $445 million credit facility provided by Taiwan’s CTBC. San Francisco-based investment firm ARCHIV led the round, with participation from Nvidia and Taiwan’s Trend Micro, among others. Headquartered in Mountain View, California, GMI Cloud provides high-performance GPU infrastructure and AI inference services, with contracted annual recurring revenue (ARR) growing more than ninefold since the end of 2025. The proceeds will be used to expand GPU supply capacity across the U.S. and Asia-Pacific regions, including Taiwan, and to enhance inference service capabilities.
Key Elements
KB Investment, the venture capital arm of South Korea’s KB Financial Group, along with DSC Investment, Kyobo Life Insurance, and KT, have invested in GMI Cloud, a U.S.-based AI-native cloud company, alongside Nvidia. The move is seen as an effort to secure a foothold in the core infrastructure segment of the AI value chain amid surging demand for GPU computing driven by the proliferation of generative AI.
According to investment banking sources on the 1st, GMI Cloud has completed a Series B funding round totaling $668 million (approximately ₩910 billion). Of this amount, equity investment accounts for $223 million (approximately ₩300 billion), with Taiwan’s CTBC providing a $445 million (approximately ₩610 billion) credit facility.
The round was led by ARCHIV, a San Francisco-based investment firm specializing in AI and robotics. Nvidia, the global leader in AI semiconductors, participated as an investor, alongside Asia-Pacific investors including Taiwan’s Trend Micro. From South Korea, KB Investment, DSC Investment, Kyobo Life Insurance, and KT joined the round, underscoring the growing presence of South Korean capital in global AI infrastructure deals.
Headquartered in Mountain View, California, GMI Cloud provides high-performance GPU infrastructure and AI inference services. The company has experienced rapid growth amid surging demand for GPU computing. Contracted annual recurring revenue (ARR) has grown more than ninefold since the end of 2025, while ARR from actively operating services increased more than 4.5 times over the same period. The inference platform processes approximately 4 trillion tokens per week. Key customers include Fireworks, Higgsfield, Nous Research, OpenRouter, Reflection, Cartesia, Trend Micro, and Utopai Studios.
The proceeds will be deployed to expand GPU supply capacity across the U.S. and Asia-Pacific regions, including Taiwan. GMI Cloud plans to build out infrastructure centered on its Taiwan AI factory—announced in 2025 as its first Asian foothold—and its Japan sovereign AI initiative unveiled earlier this year. The funding will also be used to recruit key talent for inference service enhancement and business expansion.
A Dual Strategy Bridging the U.S. and Asia
GMI Cloud’s core competitive advantage lies in its simultaneous pursuit of both the U.S. and Asian markets. U.S. AI companies and major cloud providers need local computing resources to deliver fast, reliable services to Asian users. Conversely, Asian enterprises face regulatory environments that make it difficult to move data across borders, requiring infrastructure that allows them to operate AI services within their own countries.
GMI Cloud addresses both demands by enabling users to deploy workloads in their preferred regions based on location, data, and regulatory requirements. Supply chain stability is another key strength. The company works closely with Taiwanese manufacturers that produce the majority of the world’s AI servers, earning a reputation for predictable timelines from order placement to deployment.
“Delivery timelines in AI infrastructure are promises to customers—they build product launches, hiring plans, and revenue forecasts on that foundation,” said Alex Yeh, CEO of GMI Cloud. “Building on our established position in the Taiwanese supply chain, we will construct reliable computing infrastructure across multiple continents and grow a product ecosystem on top of it.”
Chenyu Zhao, co-founder of Fireworks, commented: “Computing resources not secured on time are resources you cannot use. GMI Cloud stands out among partners supplying NVIDIA GB200 and GB300 NVL72 systems as a core partner with exceptional capability and high reliability.”
Investment Structure and Market Implications
A distinguishing feature of this Series B round is its hybrid structure combining equity investment with credit facilities. By providing a $445 million credit facility, Taiwan’s CTBC has enabled GMI Cloud to lower its capital financing costs for large-scale GPU infrastructure buildout. Given the massive upfront capital requirements characteristic of AI infrastructure businesses, such financial structures can be particularly advantageous for growth-stage companies.
The participation of South Korean investors demonstrates that Korean capital’s interest in AI infrastructure is expanding beyond purely financial investment into strategic dimensions. KT’s involvement, in particular, can be interpreted as an effort by the telecommunications infrastructure operator to deepen its understanding of the AI cloud ecosystem. Kyobo Life Insurance, for its part, appears to have focused on the stable cash flows of AI infrastructure from the perspective of long-term asset management typical of insurers.
The GPU computing market is being reshaped around Nvidia’s high-performance chips, including the H100, GB200, and GB300. GMI Cloud’s positioning as a partner supplying Nvidia’s latest NVL72 systems was a key factor behind this successful fundraising. Nvidia’s direct participation as an investor is also read as a signal of confidence in GMI Cloud that goes beyond a mere capital relationship, extending to supply chain stability.
AI inference demand is growing rapidly as the market shifts from the training phase to actual service operations. GMI Cloud’s processing of 4 trillion tokens per week is evidence that generative AI services have entered the commercialization stage. Investment industry observers expect the inference segment of AI infrastructure to post the highest growth rates over the coming years.
Once added, BigGo Finance appears first in Google Search Top Stories, so you get the broadest, most up-to-the-minute, and most comprehensive global financial news first.
