Charles Liang, CEO of Supermicro, asserts that “the AI revolution could surpass the industrial revolution,” a claim supported by Supermicro’s remarkable growth over the past year. The company’s revenue surged by 200% year-over-year in the third fiscal quarter, with projections indicating a further 152% growth by year-end. Speaking at VB Transform 2024, Liang highlighted the substantial wave of profitability Supermicro has experienced, backed by trends fueling investor confidence in the ongoing AI revolution.
Supermicro is strategically positioned within the AI landscape. While AI relies heavily on GPUs produced primarily by Nvidia, Supermicro complements these offerings. The company provides clusters and racks of networked, cooled, pooled AI infrastructure that meets the rising demand for customizable, high-performance solutions.
“As a Silicon Valley-based company, we collaborate closely with industry leaders like Nvidia, Intel, AMD, and Broadcom,” Liang noted. “Our engineering team partners with theirs tirelessly to develop their products. We also work closely with major software companies and key customers to design highly optimized platforms using our building block solutions, ensuring the best time to market.”
Preparing for the Future of AI
Liang emphasizes that the current AI boom is just the beginning: “We are expanding our capacity to support the global industry.”
In Silicon Valley, Supermicro is building and shipping 4,000 racks monthly, with around 1,000 featuring liquid cooling solutions. The company plans to increase this to about 2,000 liquid-cooled racks per month. Liang highlighted the necessity of liquid cooling as power constraints become more pressing. Data centers currently consume substantial energy, straining utility providers. However, he asserts that liquid cooling offers a viable solution, noting that companies like Elon Musk's xAI are utilizing Supermicro’s liquid cooling systems.
“With liquid cooling, our data center solutions reduce energy consumption by 30 to 40%, enabling customers to save on energy costs while deploying 30% more computing power within the same power budget,” Liang explained. “Our goal is to ensure that liquid cooling comprises at least 20% to 30% of our data center deployments in the next 12 months.”