Supermicro's Liquid Cooling Innovation Boosts AI Computing Power

By embracing liquid cooling, AI powerhouse Supermicro enables 30% more computing power -- with the same power budget

Charles Liang, CEO of Supermicro, predicts that "the AI revolution can be bigger than the industrial revolution," and the astonishing growth of Supermicro just in the past year seems to bear out this prediction. Supermicro revenue grew by 200% year over year in the third fiscal quarter, and it's expected to grow another 152% by the end of the year. Liang spoke at VB Transform 2024 about the massive wave of profitability the company has experienced and the trends that are underpinning investor confidence that the AI revolution, and its profitability for well-positioned companies, will continue. And Supermicro is better positioned than most companies. AI depends on the GPUs that companies like Nvidia (mostly Nvidia) are producing, but without Supermicro, you only have GPUs. Supermicro ships clusters and racks of networked, cooled and pooled AI infrastructure (including GPUs) to fuel continuous demand for its best-in-class building block solutions that let customers customize every component of their platform. "As a Silicon Valley-based company, we are able to work with all the technology leaders from Nvidia, Intel, AMD and Broadcom and so on, so our engineering team works with their team from early morning to midnight if their product needs to be developed," Liang said. "And we work with all the major software companies very closely and all our leading customers to brainstorm and design the most optimized platforms for them, based on our building block solution, with the best time to market." Preparing for the future of AI "This AI boom, I believe that's just at a beginning stage," Liang said. "So, we are preparing capacity to support the whole industry around the world." In Silicon Valley alone, they're building and shipping 4,000 racks per month. About 1,000 of that number offers a liquid cooling solution - and they're ramping up right now to ship about 2K liquid-cooled racks a month. Liquid cooling will be an essential factor going forward, Liang added, now that power constraints are rearing their ugly head. Right now, data centers consume massive amounts of power, and utilities are having a hard time keeping up. But liquid cooling is the answer moving forward, Liang says. In fact, Elon Musk is using Supermicro's liquid cooling solution for xAI. "With liquid cooling, our data center solutions consume 30 to 40% less energy, so that not just saves customers energy cost, but also enables customers to deploy 30% more computing power with the same power budget," he explained. Our goal is to make liquid cooling at least 20% to 30% of the whole data center deployment in next 12 months."

Software-led solutions a focus during Supermicro event - SiliconANGLE

Amid the artificial intelligence era, innovations in storage solutions are reshaping how businesses manage and utilize data. Interest in high-performance computing, media and entertainment, hyperconverged infrastructure, software-led solutions and AI has never been higher. That's an area in which Super Micro Computer Inc. seeks to capitalize. The upcoming Supermicro Open Storage Summit event will be a chance to dig into what software-led solutions will apply to an organization's projects now and in the future, according to Rob Strechay, managing director and principal analyst with theCUBE Research. "The Supermicro Open Storage Summit is a fantastic opportunity to not only understand the products that are being built with incredibly innovative partners of Supermicro, but to hear the perspectives of tens of thousands of customers on how data platforms are designed for AI, software-defined storage, HCI, HPC, networking for AI, data strategies and storage strategies," Strechay said. Exploring Supermicro's ecosystem will be of key focus during the Supermicro Open Storage Summit event. TheCUBE's coverage will explore how the company seeks to redefine infrastructure and software-led solutions for the future of enterprise AI. Those subjects and more will be explored, including exclusive details on Supermicro products, workflows and solutions. Join theCUBE from August 13 to 29 for our coverage of the Supermicro Open Storage Summit event live on theCUBE, SiliconANGLE Media's livestreaming studio. (* Disclosure below.) The following is a snapshot of Supermicro's ecosystem as it redefines infrastructure and software-led solutions for the future of enterprise AI. The company's thought leadership spans high-performance computing, media and entertainment, hyperconverged infrastructure and AI, setting new standards in the industry. In late April, Supermicro reported net sales of $3.85 billion, a substantial increase from the $1.28 billion reported in the same quarter the previous year. The market strength can be attributed to strong demand for AI rack scale PnP solutions, along with other measures, according to Charles Liang, president and CEO of the company. "[The AI rack scale PnP solutions], along with our team's ability to develop innovative DLC designs, enabled us to expand our market leadership in AI infrastructure. As new solutions ramp, including fully production-ready DLC, we expect to continue gaining market share," Liang wrote in a statement. The company also raised its fiscal year 2024 revenue outlook from $14.3 to $14.7 billion to a new range of $14.7 to $15.1 billion. It has certainly been a notable run for Supermicro in recent months, with some pointing to the company's use cases from ecosystem partners, particularly Nvidia Corp. The company is doing well and has essentially doubled all of its key performance indicators, according to Holger Mueller of Constellation Research Inc. "Supermicro is doing really well, basically doubling all of its key performance indicators, from revenue to profit to earnings per share," the analyst said. "But today's investor reaction is a sharp reminder that even in the middle of a boom, you need to manage to guidance. Charles Liang and team will want to improve on that in the next quarter, just as much as the company's shareholders do." In recent months the company has sought to further outline its products, workflows and software-led solutions. That includes a June announcement tied to a ready-to-deploy liquid-cooled AI data center. The data center is designed for cloud-native solutions, intended to accelerate generative AI adoption for enterprises across industries. That's done via the company's AI SuperClusters, which are optimized for the Nvidia AI Enterprise software platform. "The result is that we can reduce the time to delivery of our liquid-cooled or air-cooled turnkey clusters with Nvidia HGX H100 and H200, as well as the upcoming B100, B200, and GB200 solutions," Liang wrote in a statement. "From cold plates to CDUs to cooling towers, our rack-scale total liquid cooling solutions can reduce ongoing data center power usage by up to 40%." Later that same month, the company announced that it would be adding three new manufacturing facilities in Silicon Valley and globally, citing demand for liquid-cooled data centers. The facilities would be part of the new liquid-cooled ecosystem, according to the company. "Supermicro is developing building block liquid-cooled solutions for AI factories and the HPC market," Liang wrote in a statement. "We anticipate that liquid-cooled data centers will grow from historically less than 1% to an expected 15% and up to 30% of all data center installations in the next two years. This expansion positions us to capture the majority share of that growth." The company's results have been impressive, but it will still need to work on the profitability side of its business, according to Mueller. That said, the company still has good news in store for investors. "The driving force behind this growth is AI, as Supermicro provides some of the most reliable computing platforms for these workloads thanks to its partnership with Nvidia," Mueller said after the company's second-quarter financial results release. "The good news for investors is that this trend is unlikely to change anytime soon." Diving into the details of Supermicro products as infrastructure is redefined to meet generative AI demands will be the main focus of the Supermicro Open Storage Summit event. TheCUBE will provide full coverage of the event while providing analyst-driven commentary. You can follow theCUBE's wall-to-wall coverage for firsthand insights. During the Supermicro Open Storage Summit event, theCUBE analysts will talk with industry professionals about Supermicro's open storage solutions. (* Disclosure: TheCUBE is a paid media partner for the Supermicro Open Storage Summit event. Neither Super Micro Computer Inc., the sponsor of theCUBE's event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)