0:00
0:00
Nvidia GPUs are so hard to get that rich venture capitalists are buying them for the startups they invest in
Rich VCs Purchase GPUs and Establish Andromeda Cluster AI Cloud Service
The high demand and limited supply of Nvidia GPUs, essential for training big models in generative AI tools like ChatGPT, have created a challenging landscape for AI startups. While big tech companies with substantial resources have been able to navigate the scarcity, smaller startups face hurdles due to soaring prices and shortages. In response, some venture capitalists (VCs) are taking unconventional measures to assist these startups.
Nvidia GPUs, particularly the A100 and H100 chips, are critical components fueling today's thriving AI ecosystem. The power these chips provide is instrumental in the development of advanced generative AI services. However, the unprecedented demand for GPUs has outpaced the manufacturing capacity of Nvidia and its partner, TSMC. While Nvidia has experienced significant market growth, with nearly $200 billion added to its market cap after announcing higher revenue guidance, the shortage of GPUs has negatively impacted AI startups, leading to skyrocketing prices and supply gaps.
This discrepancy gives an advantage to established Big Tech companies, such as Microsoft, OpenAI, Google, and even Adobe, who possess the financial means to invest in training large foundation models with vast amounts of data. To compete in the AI landscape, startups often need to train their own models using GPUs, a costly endeavor that many cannot afford or cannot secure due to the shortages.
Even tech giant Microsoft has faced challenges, with reports indicating it is rationing internal access to GPUs to conserve processing power for AI-powered applications like Bing chatbot and AI Microsoft Office tools.
In light of these difficulties, certain venture capitalists have taken proactive measures to address the issue. Notably, Nat Friedman, former CEO of GitHub, and Daniel Gross, a prominent investor in successful startups like GitHub and Uber, have stepped forward. They have purchased thousands of GPUs and established their own AI cloud service, known as the Andromeda Cluster.
The Andromeda Cluster comprises an impressive 2,512 H100 GPUs, capable of training a 65 billion parameter AI model within approximately 10 days, according to the VCs involved. While the offering is exclusively available to startups backed by Friedman and Gross, the initiative has garnered praise for its innovative approach.
The significance of individual investors taking greater strides to support compute-intensive startups has not gone unnoticed. Jack Clark, co-founder of AI startup Anthropic, expressed his admiration for the project on Twitter, highlighting the meaningful contribution made by these VCs.
As the demand for GPUs persists and AI startups seek to overcome the challenges posed by shortages and rising costs, the actions of these venture capitalists provide a glimmer of hope and support for the flourishing AI ecosystem.
Nvidia GPUs, particularly the A100 and H100 chips, are critical components fueling today's thriving AI ecosystem. The power these chips provide is instrumental in the development of advanced generative AI services. However, the unprecedented demand for GPUs has outpaced the manufacturing capacity of Nvidia and its partner, TSMC. While Nvidia has experienced significant market growth, with nearly $200 billion added to its market cap after announcing higher revenue guidance, the shortage of GPUs has negatively impacted AI startups, leading to skyrocketing prices and supply gaps.
This discrepancy gives an advantage to established Big Tech companies, such as Microsoft, OpenAI, Google, and even Adobe, who possess the financial means to invest in training large foundation models with vast amounts of data. To compete in the AI landscape, startups often need to train their own models using GPUs, a costly endeavor that many cannot afford or cannot secure due to the shortages.
Even tech giant Microsoft has faced challenges, with reports indicating it is rationing internal access to GPUs to conserve processing power for AI-powered applications like Bing chatbot and AI Microsoft Office tools.
In light of these difficulties, certain venture capitalists have taken proactive measures to address the issue. Notably, Nat Friedman, former CEO of GitHub, and Daniel Gross, a prominent investor in successful startups like GitHub and Uber, have stepped forward. They have purchased thousands of GPUs and established their own AI cloud service, known as the Andromeda Cluster.
The Andromeda Cluster comprises an impressive 2,512 H100 GPUs, capable of training a 65 billion parameter AI model within approximately 10 days, according to the VCs involved. While the offering is exclusively available to startups backed by Friedman and Gross, the initiative has garnered praise for its innovative approach.
The significance of individual investors taking greater strides to support compute-intensive startups has not gone unnoticed. Jack Clark, co-founder of AI startup Anthropic, expressed his admiration for the project on Twitter, highlighting the meaningful contribution made by these VCs.
As the demand for GPUs persists and AI startups seek to overcome the challenges posed by shortages and rising costs, the actions of these venture capitalists provide a glimmer of hope and support for the flourishing AI ecosystem.
