The GenAI Datacenter Squeeze Is Here

By Doug Eadline

February 1, 2024

The immediate effect of the GenAI GPU Squeeze was to reduce availability, either direct purchase or cloud access, increase cost, and push demand through the roof.

A secondary issue has been developing over the last several years. Even though your organization secured several racks with GPUs, how will you power them, and where will you put them?

For instance, many universities have previously placed new equipment in their campus datacenter. Many of these datacenters are now “tapped out” and have no more space or power. Estimates for the current “GPU rack” range from 50 to 100 KW (kilowatts) per rack (previous estimations for “CPU racks” are 10-17 KW per rack). If you want to co-locate 4 GPU racks, a datacenter that can provide 400 KW of power may be difficult to find.

The same applies to the smaller companies that provide HPC colocation services. They are finding that current datacenters are tapped out and must search for space and power. These companies are the non-Hyperscalers that can’t stand up an entire datacenter campus.

A recent report from JLL, a real estate investment and management firm, Data Centers 2024 Global Outlook  (The report can be obtained by providing an email address), shares some interesting trends. The report explores how datacenters need to be designed, operated, and sourced to meet the evolving needs of the global economy, in particular, the power increase required for GPU-heavy GenAI clusters.

The AI-fueled growth is expected to continue into the near future. Consumers and businesses are anticipated to generate twice as much data in the next five years as all the data created over the past ten years.

In addition to the GPU computing needs, the growing demands of GenAI, datacenter storage capacity is expected to grow from 10.1 zettabytes (ZB) in 2023 to 21.0 ZB in 2027, for a five-year compound annual growth rate of 18.5%[1]. This increased storage will generate a need for more datacenters, and generative AI’s greater energy requirements – ranging from 300 to 500+ megawatts per campus – will also require more energy-efficient designs and locations. The need for more power will require datacenter operators to increase efficiency and work with local governments to find sustainable energy sources to support datacenter needs.

According to the report, AI-specialized datacenters look different than conventional facilities and may require operators to plan, design, and allocate power resources based on the type of data processed or stage of GenAI development. Regarding heat removal, the huge increase in GPUs will surpass current standards. In a typical datacenter, air-cooling typically accounts for roughly 40% of an average datacenter’s electricity use. Users, particularly hyperscalers and operators, are shifting from traditional air-based cooling methods to liquid cooling and rear-door heat exchangers. Case studies have shown that liquid cooling offers significant power reductions, as high as 90%, while improving capability and allowing more densely packed systems that, in turn, increase rack power usage.

Hyperscalers, which have been at the forefront of adopting AI and high-performance computing (HPC), have the greatest need for high-density infrastructure (See Figure). Currently, their large facilities have an estimated average density of 36kW per rack, with increased liquid cooling density and GPU hardware IDC estimates will grow at a 7.8% CAGR in the coming years to approach 50kW per rack by 2027. [2]

(Source JLL: Data Centers 2024 Global Outlook)

In an interview with Andy Cvengros, managing director of U.S. Data Center Markets for JLL, HPCwire learned many issues facing datacenter deployments.

The first and most important issue he suggests is planning. For instance, Cvengros mentioned that with power grids becoming effectively tapped out and transformers having more than three-year lead times, operators will need to innovate. 

The GPU squeeze is taking pace at the datacenter level, where a small colocation deployment of 4-5 racks will have a harder time finding a datacenter because the hyperscalers are requesting entire datacenter campuses.

According to Cvengros, all major metro areas are basically tapped out, and secondary areas, like Reno, NV, or Columbus, OH, are now prime locations for new datacenter construction. The demand is expected to continue, and new datacenters are 3.5 years out. He reiterated, “Planning is key.”

He recommends working with a company that specializes in high-performance systems for smaller HPC GPU-Cluster colocation (i.e., the University that is completely tapped out of power and space). In his experience, datacenter providers must actively track usage and availability for global opportunities to deliver any near-term colocation capabilities.

The Datacenter Squeeze is a Global Problem

The JLL report also lists the critical changes needed across the globe to address increased power usage.

  • In Europe, one-third of the grid infrastructure is over 40 years old, requiring an estimated €584 billion of investment by 2030 to meet the European Union’s green goals.
  • In the United States, meeting energy transition goals to upgrade the grid and feed more renewable energy into the power supply will require an estimated $2 trillion.
  • The rapid growth of datacenters is also putting pressure on limited energy resources in many countries. In Singapore, for example, the government enacted a moratorium to temporarily halt construction in certain regions to carefully review new datacenter proposals and ensure alignment with the country’s sustainability goals.

The global GenAI energy “request” (or demand) presents opportunities and challenges to the datacenter sector. GenAI needs power on a scale not seen in the past. Finding GPUs for HPC is only half the problem; where to plug them in may become a bigger challenge.


[1] IDC, Revelations in the Global StorageSphere, July 2023

[2] July 2023, IDC #AP50326223, Asia/Pacific (Excluding Japan) DC Deployment Model and Spend Forecast, 2H22: 2022–20a27

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

MLPerf Training 4.0 – Nvidia Still King; Power and LLM Fine Tuning Added

June 12, 2024

There are really two stories packaged in the most recent MLPerf  Training 4.0 results, released today. The first, of course, is the results. Nvidia (currently king of accelerated computing) wins again, sweeping all nine Read more…

Highlights from GlobusWorld 2024: The Conference for Reimagining Research IT

June 11, 2024

The Globus user conference, now in its 22nd year, brought together over 180 researchers, system administrators, developers, and IT leaders from 55 top research computing centers, national labs, federal agencies, and univ Read more…

Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study

June 10, 2024

Nvidia had an explosive 2023 in data-center GPU shipments, which totaled roughly 3.76 million units, according to a study conducted by semiconductor analyst firm TechInsights. Nvidia's GPU shipments in 2023 grew by more Read more…

Weekly Wire Roundup: June 2-June 7, 2024

June 8, 2024

Computex (and Jensen Huang) gave us an extra day of news this week, compensating for last week's shorter, holiday-driven news cycle. On Sunday ahead of the official start of Computex, Nvidia's CEO Jensen Huang deliver Read more…

ASC24 Expert Perspective: Dongarra, Hoefler, Yong Lin

June 7, 2024

One of the great things about being at an ASC (Asia Supercomputer Community) cluster competition is getting the chance to interview various industry experts and learning more about the various challenges the students are Read more…

HPC and Climate: Coastal Hurricanes Around the World Are Intensifying Faster

June 6, 2024

Hurricanes are among the world's most destructive natural hazards. Their environment shapes their ability to deliver damage; conditions like warm ocean waters, guiding winds, and atmospheric moisture can all dictate stor Read more…

MLPerf Training 4.0 – Nvidia Still King; Power and LLM Fine Tuning Added

June 12, 2024

There are really two stories packaged in the most recent MLPerf  Training 4.0 results, released today. The first, of course, is the results. Nvidia (currently Read more…

Highlights from GlobusWorld 2024: The Conference for Reimagining Research IT

June 11, 2024

The Globus user conference, now in its 22nd year, brought together over 180 researchers, system administrators, developers, and IT leaders from 55 top research Read more…

Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study

June 10, 2024

Nvidia had an explosive 2023 in data-center GPU shipments, which totaled roughly 3.76 million units, according to a study conducted by semiconductor analyst fir Read more…

ASC24 Expert Perspective: Dongarra, Hoefler, Yong Lin

June 7, 2024

One of the great things about being at an ASC (Asia Supercomputer Community) cluster competition is getting the chance to interview various industry experts and Read more…

HPC and Climate: Coastal Hurricanes Around the World Are Intensifying Faster

June 6, 2024

Hurricanes are among the world's most destructive natural hazards. Their environment shapes their ability to deliver damage; conditions like warm ocean waters, Read more…

ASC24: The Battle, The Apps, and The Competitors

June 5, 2024

The ASC24 (Asia Supercomputer Community) Student Cluster Competition was one for the ages. More than 350 university teams worked for months in the preliminary competition to earn one of the 25 final competition slots. The winning teams... Read more…

Computex 2024: Nvidia, AMD Push GPUs; Intel Revs Up x86 Power Efficiency

June 5, 2024

"The days of millions of GPU data centers are coming," said Nvidia CEO Jensen Huang during a keynote at Computex. Huang's predictions are becoming bolder and bo Read more…

Using AI and Robots to Advance Science

June 4, 2024

Even though we invented it, humans can be pretty bad at science. We need to eat and sleep, we sometimes let our emotions regulate our behavior, and our bodies a Read more…

Atos Outlines Plans to Get Acquired, and a Path Forward

May 21, 2024

Atos – via its subsidiary Eviden – is the second major supercomputer maker outside of HPE, while others have largely dropped out. The lack of integrators and Atos' financial turmoil have the HPC market worried. If Atos goes under, HPE will be the only major option for building large-scale systems. Read more…

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Everyone Except Nvidia Forms Ultra Accelerator Link (UALink) Consortium

May 30, 2024

Consider the GPU. An island of SIMD greatness that makes light work of matrix math. Originally designed to rapidly paint dots on a computer monitor, it was then Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

Leading Solution Providers

Contributors

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

The NASA Black Hole Plunge

May 7, 2024

We have all thought about it. No one has done it, but now, thanks to HPC, we see what it looks like. Hold on to your feet because NASA has released videos of wh Read more…

GenAI Having Major Impact on Data Culture, Survey Says

February 21, 2024

While 2023 was the year of GenAI, the adoption rates for GenAI did not match expectations. Most organizations are continuing to invest in GenAI but are yet to Read more…

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's l Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

Intel’s Next-gen Falcon Shores Coming Out in Late 2025 

April 30, 2024

It's a long wait for customers hanging on for Intel's next-generation GPU, Falcon Shores, which will be released in late 2025.  "Then we have a rich, a very Read more…

Intel Plans Falcon Shores 2 GPU Supercomputing Chip for 2026  

August 8, 2023

Intel is planning to onboard a new version of the Falcon Shores chip in 2026, which is code-named Falcon Shores 2. The new product was announced by CEO Pat Gel Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire