| Ticker | Status | Jurisdiction | Filing Date | CP Start | CP End | CP Loss | Deadline |
|---|
| Ticker | Case Name | Status | CP Start | CP End | Deadline | Settlement Amt |
|---|
| Ticker | Name | Date | Analyst Firm | Up/Down | Target ($) | Rating Change | Rating Current |
|---|
As AI technology advances at breakneck speed, the power demands of data centers are reaching unprecedented heights. GPU-world titans, NVIDIA Corp (NASDAQ:NVDA) and Advanced Micro Devices, Inc. (NASDAQ:AMD), are pushing the boundaries with their cutting-edge AI accelerators.
This surge in power needs is now necessitating significant upgrades in thermal management solutions. A prime example of this is the anticipated shift from air to liquid cooling systems.
The rise in AI-driven power demand isn’t just a technical challenge — it’s a golden opportunity for companies like Vertiv Holdings Co (NYSE:VRT). Vertiv is poised to revolutionize thermal management solutions with its in-row precision cooling solution designed for high-density applications.
According to JPMorgan Forwarding Research on AI Datacenters, data center power consumption is projected to exceed 3% of global electricity usage by 2028. This is up from approximately 2% in 2022.
This surge underscores the critical role of power-efficient technologies and regulatory measures in achieving net zero emission (NZE) goals by 2050. Companies that stand to benefit from this trend include: Delta, thermal vendors such as AVC and Auras, and power management firms such as Vertiv.
AI silicon upgrades are pivotal to the rising power demands in data centers. Unlike the longer upgrade cycles of CPUs, GPU vendors are accelerating their product roadmaps to meet the increasing requirements of large language model (LLM) training and inference. This aggressive pace is evident in the power leaps between product generations.
Nvidia exemplifies this trend. The company’s GPUs have seen substantial increases in thermal design power (TDP).
For instance, the TDP of Nvidia's P100 GPU, launched in 2016, was 250W. By 2020, the A100 GPU’s TDP had risen to 400W, reflecting a compound annual growth rate (CAGR) of about 15%. Post-2022, spurred by the AI inflection point marked by ChatGPT’s launch, Nvidia has expedited its product releases, planning new launches annually. This aggressive schedule included:
These are anticipated to result in a ~30% power CAGR for GPU silicon.
Also Read: Nvidia’s Success May Propel These 6 AI-Related Stocks To New Heights
AMD is following a similar trajectory with its high-end data center GPUs. The TDP of AMD's MI250, launched in the fourth quarter of 2021, stands at 560W. This is expected to nearly triple to 1600W with the forthcoming MI375/400 series, anticipated around the fourth quarter or 2025. AMD’s GPU power requirements are anticipated to grow at a ~30% CAGR.
As power consumption soars, efficient thermal management becomes crucial. JPMorgan has identified Vertiv, a leader in this space, as being well-positioned to capitalize on the shift.
The company’s expertise in thermal management is essential as data centers transition from traditional air cooling to more sophisticated liquid cooling systems. The escalating power demands of AI accelerators and the corresponding advancements in thermal management highlight significant investment opportunities.
Nvidia and AMD's rapid GPU development underscores the urgent need for efficient power and cooling solutions, with companies like Vertiv poised to benefit.
Read Next: AI-Driven Rally Set To Expand To Power, Commodities, Utilities: ‘It’s Not Just About Nvidia Anymore’
Photo: Vertiv SmartRow 2, courtesy Vertiv
