Super Micro Computer Inc. (trading as ‘Supermicro’) has announced it is shipping over 100,000 graphics processing units (GPUs) per quarter, coinciding with the introduction of new liquid cooling products. This announcement led to a 14% increase in the Nasdaq-listed company’s share price.

The growing demand for generative artificial intelligence (AI) has increased the need for hardware capable of processing large data sets. Supermicro, a manufacturer of servers featuring AI chips like those from Nvidia, has capitalised on this trend with its recent developments in liquid cooling technology, which has been deployed in large AI factories.

This rise in the company’s stock could potentially add more than $3bn to its market value, reported Reuters. Supermicro’s liquid cooling technology has been noted for its efficiency in reducing power consumption compared to air-cooling methods in data centres.

Supermicro stock on the up and up amid AI chip boom

Supermicro’s stock increase comes after a period of decline, partly due to the disclosure in August of a short position by the hedge fund Hindenburg Research, with the hedge fund alleging accounting manipulations and self-dealing at the chip business. Ortex, a market research firm, estimates that short interest in Supermicro’s shares accounts for over 20% of the company’s free float, valued at around $3.59bn. Ortex indicated that it is uncertain if the stock rise is related to short sellers covering their positions but noted short-term losses for those holding short positions.

Supermicro’s shares had dropped by over 9% in recent weeks, but they are up by more than 66% this year, reflecting broader trends tied to AI investments.

The company’s newly introduced liquid cooling solutions are capable of supporting up to 96 Nvidia B200 chips per rack. Gadjo Sevilla, a senior AI analyst at eMarketer, highlighted that these products offer significant efficiencies in energy and space use, positioning them as key in the growing AI infrastructure market.

Supermicro has released a full suite of liquid cooling products designed to reduce power and infrastructure costs in data centres. These include Coolant Distribution Units, cold plates, Coolant Distribution Manifolds, and management software. This solution is available directly from Supermicro and is aimed at supporting AI and high-performance computing workloads.

The company’s servers, featuring dual CPUs and eight Nvidia HGX GPUs, have been designed for large AI factories and cloud service providers, providing more efficient cooling and improved computing density.

“Supermicro continues to innovate, delivering full data centre plug-and-play rack scale liquid cooling solutions,” extolled Supermicro’s CEO and president Charles Liang. “Our complete liquid cooling solutions, including SuperCloud Composer for the entire life-cycle management of all components, are now cooling massive, state-of-the-art AI factories, reducing costs and improving performance.”

Read more: Samsung set to report Q3 2024 profit gains despite lagging in AI market