Supermicro has seen substantial growth over the past two years, driven by rising demand for high-performance GPU systems used in AI training and inference
In sum – what you need to know:
AI demand is driving global expansion – Supermicro is building up operations in Europe and Asia to support rapid international growth in AI infrastructure.
Liquid cooling – The company shipped 4,000 direct liquid cooling racks in 2024, enabling significant energy savings.
AI is part of Supermicro’s own R&D – Nearly 40% of the company’s firmware and software is now AI-assisted, accelerating innovation while scaling operations.
San Jose-based company Supermicro is ramping up its investments in energy-efficient AI infrastructure and expanding its global manufacturing capabilities to meet accelerating demand for advanced computing systems.
Speaking at the RAISE Summit 2025, company CEO Charles Liang outlined Supermicros’s strategic priorities amid a booming market for AI workloads, including growing its footprint in Europe, deepening support for green data center solutions, and navigating increasing competition in the server market.
The executive noted that Supermicro has seen substantial growth over the past two years, driven by rising demand for high-performance GPU systems used in AI training and inference. Liang believes this trend is only just beginning.
“I personally believe this AI boom trend will continue for many years to come,” he said during a fireside chat at the event.
While the initial wave of demand originated in the U.S., Liang said AI infrastructure adoption is accelerating in the Middle East and Europe, prompting Supermicro to scale its international operations. “Everyone see that demand grow to Middle East, grow to Europe very strong and very fast now,” he noted.
In response to this growing demand, Supermicro is expanding its European presence with local assembly, service support, and logistics capabilities.
The company’s modular “building block” design and vertically integrated model have allowed it to move quickly and customize offerings for diverse workloads. Still, Liang acknowledged growing pricing pressure as more firms enter the AI server market and volumes surge.
“When the volume is huge, when there are lots of competition, for sure, they always have a cost of price pressure,” he said. While competition can erode margins, Liang believes it also forces faster innovation. “The more competition, [makes] the industry grow even faster, even better.”
He also explained that a key differentiator for Supermicro is its leadership in liquid cooling—a technology increasingly necessary for dense, high-wattage AI systems. In 2024 alone, the company shipped 4,000 racks with direct liquid cooling, Liang said.
These systems can reduce data center energy consumption by up to 40%, according to customer feedback. “We work together to save them up to 30%, or some customer even save up to 40% of energy power,” Liang said. He added that Supermicro spent three years refining its liquid cooling technology to improve reliability and simplify deployment.
Supermicro is also adopting AI tools in its own operations, especially in hardware design and firmware development. According to Liang, nearly 40% of the company’s software and firmware components are now designed with the help of AI, allowing the company to accelerate development cycles without reducing headcount. “Today, up to 40% of our firmware, software, are designed by AI,” he said.
Looking ahead, Supermicro plans to evolve beyond a hardware vendor to offer integrated AI data center solutions. “We start to extend our business scope to data center, infrastructure, total solution,” Liang said, citing demand from hyperscalers and enterprises that want faster deployment of full-stack AI systems.
While geopolitical and economic uncertainty could slow the market in the short term, Liang remains optimistic about AI’s central role in global competitiveness.
