GMI Cloud is building a new GPU infrastructure

The boom in artificial intelligence has created a huge demand for graphics processing units (GPUs). As many enterprises are looking to integrate artificial intelligence technologies into their systems, GPU infrastructure providers help them get access to the chips they need.

Cloud infrastructure

GMI Cloud, a San Jose-based startup that provides GPU cloud infrastructure, raised 82 million dollars in a Series A led by Headline Asia and strategic investors such as Banpu, a Thailand-based energy company, and Wistron, a Taiwan-based electronics company. Banpu will offer GMI Cloud capacity, while Wistron will develop products with the startup. The strategic partnership expands the company's ability to meet the growing global demand for GPUs, GMI Cloud founder and CEO Alex Ye told TechCrunch. The investment round, which brings the total capital raised to 93 mln dollars, consists of 15 mln dollars in equity and 67 mln dollars in debt financing.

Company launch

The company, which launched in 2022, started as a data center focused on providing Bitcoin compute node services. In an exclusive interview with TechCrunch, Ye said he noticed a growing demand for GPU computing power from investors and customers, and in response, shifted the company's focus to cloud-based artificial intelligence infrastructure.

Ye said the two-year-old startup already serves dozens of customers, including those in healthcare, research and telecommunications. The corporation plans to use the funding to build a new data center in Colorado. The facility will be needed to expand the company's capacity in North America and will complement existing data centers in Taiwan, Thailand and Malaysia. The company also plans to expand to 60-70 employees by the end of the year. The startup now has 35 employees in Asia and 18 in the US.

According to a recent McKinsey report, artificial intelligence could bring about 13 trillion dollars to the economy by 2030, of which the industrial sector will account for about 1 trillion dollars. According to a report by Grand View Research, the global artificial intelligence market will reach 1.8 trillion dollars by 2030.

Cloud processor vendors

GMI Cloud is not the only cloud GPU vendor. It competes with Coreweave, Nebius, Google Cloud's Vertex AI and major tech companies. Ye told TechCrunch that GMI Cloud differentiates itself from competitors with features such as customizable private cloud services and native support for Nvidia NIM, which simplifies integration with Nvidia hardware and software. He also noted that the company has a team of top-notch artificial intelligence engineers and high-performance computing (HPC) experts with experience in GoogleX, Alibaba Cloud and Supermicro.

“Our team has more than 20 years of experience in artificial intelligence and high-performance computing,” said Ye, who has 33 patents in artificial intelligence and extensive experience in building large-scale distributed systems. “GMI also offers professional artificial intelligence consulting services, helping enterprises train, customize and scale models, which is rarely offered by competitors.”

Cost-effective solution

GMI offers a cost-effective solution that provides optimized performance and resource management over competitors, supporting enterprises with end-to-end solutions from GPU hardware to artificial intelligence applications, Ye emphasized.

“In addition, we have a significant supply chain advantage by purchasing directly from manufacturers, allowing us to maintain a cost-effective and highly reliable supply chain,” Ye continued. “In addition, GMI is the only cloud service provider in Taiwan certified by Nvidia under the NCP/NPN program, further strengthening our competitive advantage in providing premium cloud services.”