Nvidia H100 GPU Capacity Increasing, Usage Prices Could Get Cheaper  

By Agam Shah

November 7, 2023

It sure feels like the long lines to use Nvidia’s GPUs could get shorter in the coming months. 

A flurry of companies – large and small — in the last few months have reported receiving delivery of thousands of H100 GPUs. 

With that, the lines to use H100 GPUs in the cloud could get shorter. Datacenter providers and former Bitcoin mining companies are opening data centers with clusters of H100s. 

Those companies promise to offer H100 GPU computing at a fraction of the cost of large cloud providers, which are charging a premium for VMs accelerated by H100 GPUs.  

Amazon recently announced it is taking reservations for H100 GPUs on its website for one to 14 days in anticipation of handling surges in future demand.  

The normalizing supply is also helping companies put AI plans into action. HPCwire contacted Nvidia on whether the shortages have been resolved, but the company declined to comment. 

A few months ago, Tesla CEO Elon Musk noted the GPU shortage and was waiting to receive Nvidia GPUs to grow AI capacity. Tesla activated a cluster of 10,000 H100 GPUs, and it was deployed at a blistering pace, Musk said.  

“Training is the fundamental limiting factor on progress with full self-driving and vehicle autonomy,” Musk said on an earnings call last month. 

Nvidia has been rationing the H100s with shortages and evaluating customers based on the size of the AI models, available infrastructure, and customer profile.  

The chip maker was prioritizing customers with deep pockets across various sectors. Nvidia was also selecting customers that positively reflected the best use of its GPUs. Tesla fits the profile of being a top customer with a well-defined AI model. 

Nvidia also provides a clearer timeline on when customers will receive H100 GPUs. Datacenter provider Applied Digital purchased 34,000 H100 GPUs, of which 26,000 will be deployed by April, with an additional 8,000 coming after that. 

Applied Digital in September started receiving orders for GPUs it placed back in May. In September, it received 1,024 GPUs, and CEO Wes Cummins estimated that number doubled in October and the number could get “significantly bigger in November and December and into January.”  

The company also broke ground on a new high-performance computing datacenter in Ellendale, North Dakota, which can host 50,000 of Nvidia’s H100 SXM class GPUs in a compute cluster. 

Nvidia is also supplying smaller customers capable of signing purchase orders. 

Iris Energy, which is creating an AI computing infrastructure, paid about $10 million for 248 Nvidia GPUs. The delivery is expected in the coming months. The company said in a recent filing with the U.S. Securities and Exchange Commission last month.

The Nvidia H100 PCIe GPU

The H100s will kickstart a new AI business for Iris Energy, which earlier established itself as a cryptocurrency miner. The Australian company sees a larger generative AI market and is transforming its bitcoin mining data centers in the U.S. and Canada to also do AI computing. 

Voltage Park, a startup funded by billionaire Jed McCaleb, received a large supply of 24,000 Nvidia H100 GPUs, which are distributed across clusters in Washington state, Texas, and Virginia. 

“The order was placed around April 2023. It was clear to the team … that a crunch was coming, so they placed a very big order early to guarantee there’d be enough supply,” said Voltage Park Eric Park in an email. 

The GPUs already serve AI company Imbue, which seemingly has close ties to Voltage Park. Both Nvidia and Astera Institute – which Jed McCaleb founded – are among the investors in Imbue. 

Voltage Park also aims to make cheap H100 GPU computing capacity available to customers by supporting an initiative called FLOP Auction. It is like eBay for AI computing on H100, with the highest bidder receiving server time. 

“This is an early project looking into an auction mechanism for access to Voltage Park’s H100 clusters,” Park said, adding, “The idea is that by auctioning time as opposed to contracts, this could help to guarantee market price and a transparent process.”

The current winning bid on a FLOP Auction for a cluster of H100 per hour is cheaper than the per-hour price of V100 – which was released in 2017 – on Google Cloud. The wins are based on contract size and commitment. 

The H100 mania also has given a whole new meaning to the value of computer hardware, with the GPUs becoming collateral for new businesses.

Crusoe Energy, a former crypto miner, secured financing of $200 million with 20,000 H100s as collateral. Similarly, GPU computing provider CoreWeave has taken financing worth $2.4 billion with H100s as collateral. 

Crusoe’s GPU capacity will become available in the first quarter of next year. The company already offers clusters running on older Nvidia GPUs.

Even a cannabis company is jumping into the H100 craze. Colorado-based American Cannabis Company has agreed to merge with HyperScale Nexus Holding Corp., which wants to create a business around Nvidia’s H100 GPUs. 

The GPU shortage was exacerbated by a shortage of CoWoS packaging, which brings the memory and chip together. Nvidia was helping TSMC find new partners and suppliers to resolve shortages.  

Looking ahead, other factors could unexpectedly end shortages and speed up Nvidia’s GPU shipments. 

The U.S. government recently imposed restrictions on Nvidia’s shipment of A800, H800, and L40S series GPUs to Chinese companies. That could open CoWoS packaging and free up manufacturing capacity to make more H100 chips for the U.S. and other markets. 

Instead of outright shipping GPUs to customers, Nvidia also expands H100 capacity by renting GPUs via cloud providers. The GPU servers deployed by cloud providers are replicas of the DGX reference model, which includes eight H100 GPUs.  

“The shift might also stoke Nvidia’s enthusiasm in championing its DGX cloud subscription and leasing model, anchored by the A100 or H100 AI servers, or even broadening its horizons to include more diverse cloud services (the L40S, among others),” market research firm TrendForce said in a research note released in October.

Oracle, in September announced the general availability of its Nvidia H100 bare-metal computing service.  

Google, in late August, announced the general availability of its A3 supercomputer, which can host up to 26,000 H100 GPUs depending on computing needs.  

Microsoft also increased its GPU capacity, which provided better utilization of its AI services, the company’s CFO Amy Hood said on a recent earnings call.  

The Azure cloud service is built around Nvidia’s GPUs. Microsoft has proclaimed its Azure supercomputer as one of the fastest server systems in the world. 

The company may provide updates on its GPU supplies during the next earnings call on November 21. 

Despite red-hot demand for H100, Nvidia recently announced an updated roadmap with new GPUs planned for the next three years.  

 

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