Curious about today's AI digest?ai-tldr.dev

Daily Digest

Pomegra Startups

GMI Cloud raises $223M Series B with Nvidia in 2026

GMI Cloud (US) raised $223M, with Nvidia participating, for GPU cloud infrastructure for AI.

FundingAIMAJOR4 min read
GMI Cloud raises $223M Series B with Nvidia in 2026

GMI Cloud closed a $223M Series B led by ARCHIV with Nvidia participating, part of a $668M package that includes a $445M CTBC credit facility.

Key Takeaways

  • GMI Cloud raised $223M in Series B equity on October 1, 2026, led by ARCHIV, with Nvidia participating.
  • The equity sits beside a $445M credit facility led by CTBC, making the total package $668M.
  • Contracted annual recurring revenue exceeds $600M, 9x the end-2025 level. Valuation is undisclosed.

Lead

GMI Cloud, the Mountain View, California GPU cloud provider, announced on October 1, 2026 that it raised $223M in Series B equity, led by ARCHIV, a San Francisco firm focused on AI and robotics. Nvidia took part in the round. A $445M credit facility led by Taiwan's CTBC brings the total financing to $668M. The company said it will use the money to add capacity in the United States, Taiwan and the wider Asia-Pacific region, grow its inference services and hire.

Only a third of the headline number is equity. The rest is debt, which matters for how the figure should be read.

Who Invested in the Series B?

The Series B was led by ARCHIV, with Nvidia, DSC Investment, Trend Micro, KB Investment, Kyobo Life and KT Corporation also participating. The investor list leans heavily toward Asia. Korean financial and telecom names (KB Investment, Kyobo Life, KT) sit alongside Japanese security software maker Trend Micro, which is also a GMI customer.

GMI did not disclose a valuation. Without that figure, the equity price of the round cannot be judged, and the $223M raise cannot be compared with the company's last priced round.

What Is GMI Cloud and What Does It Sell?

GMI Cloud rents Nvidia GPU capacity to AI developers and runs an inference platform on top of it. Founder and CEO Alex Yeh leads the company, which launched in 2022 as a data center business serving Bitcoin mining before shifting to AI workloads.

Its customers include Fireworks, Higgsfield, Nous Research, OpenRouter, Reflection, Cartesia, Trend Micro and Utopai Studios. The company says it processes roughly 4 trillion tokens a week. It operates in the United States, Taiwan, Southeast Asia and Japan, opened its first Asian facility in Taiwan in 2025, and has announced a sovereign AI project in Japan.

Nvidia's role is more than that of a passive investor. GMI builds its facilities on Nvidia's recommendations, and the chipmaker counts it among a small group of cloud providers meeting that standard. In effect, Nvidia is financing a customer that buys its hardware.

How Does This Round Compare With the Series A?

The Series B is far larger than GMI's $82M Series A in October 2024, which was led by Headline Asia with Banpu and Wistron participating. That earlier round was mostly debt: about $15M in equity and $67M in borrowing. The new round repeats the pattern at greater scale, with $445M of the $668M coming from lenders.

The growth numbers explain the appetite. GMI reports contracted ARR above $600M, 9x its level at the end of 2025. Live revenue, which reflects capacity actually deployed and billing, has grown 4.5x over the same period. The gap between the two is the core of the story: signed contracts are running well ahead of racks in service, and the new money is meant to close it.

What Are the Risks in a Debt-Heavy GPU Cloud?

The main risk is that GPU-backed lending ties repayment to hardware that depreciates quickly and to customers that are themselves young. CTBC's facility is described as the first GPU-backed syndicated loan in Asia-Pacific, so there is little local precedent for how such collateral performs through a downturn. Specific loan terms were not disclosed.

Contracted revenue is also not collected revenue. If customers such as AI labs and application startups cannot raise money to honor commitments, the 9x figure could narrow. Competition adds pressure: larger Nvidia-aligned providers and hyperscalers can offer capacity at scale, and GMI's differentiation rests on regional presence and inference tooling rather than on cost of capital.

What Comes Next for GMI Cloud?

The next test is conversion: turning more than $600M of contracted ARR into live, billed capacity across the US, Taiwan and Japan. Delivery speed will decide whether the debt is serviced comfortably. A further equity round with a disclosed valuation, or an expansion of the credit line, would show whether investors still price GPU clouds on growth rather than on asset value.

Outlook

GMI Cloud has $668M in new financing, $223M of it equity from ARCHIV, Nvidia and a set of Asian strategic investors. Revenue contracts have grown quickly, but the balance sheet now leans on lenders and on Nvidia's continued backing. Valuation and loan terms remain undisclosed, and those two numbers will say most about how durable the round is.

More Startup News