The artificial intelligence wave is sweeping the world, GPUs have become a hot commodity, and Nvidia's performance is booming. However, powerful GPUs require sufficient data centers to support them, which has led to huge investments in data center construction. This article will explore the hot scene of data center construction under the AI boom, as well as the fierce competition among major companies in financing and expansion, showing the booming development of data center construction in the AI era.
AI technology is sweeping across the world like a storm, causing GPUs to become the hottest commodity in the world.
In the second quarter of this year, Nvidia's performance exploded, with revenue exceeding US$30 billion! But have you ever thought about it, just having these powerful chips, but not enough data centers to house them, would be a waste of money. ?In order to meet the continued demand for this AI craze, Nvidia and their partners began to increase investment.
Just recently, Nvidia and its partners injected up to $160 million into Texas-based Applied Digital. This company is no ordinary enterprise. They focus on providing various data centers and cloud services around Nvidia's GPUs. The funding will help Applied Digital accelerate the construction of its data center in North Dakota, while also supporting their further financing plans to pay for those high accelerator fees.
At this point, you may ask, why is the construction of data centers so important? That's because today's cutting-edge GPUs are as expensive as cars. Nvidia's upcoming Blackwell chips cost as much as $30,000 to $40,000 each. In order to raise these funds, many data center operators began to use GPUs as collateral to borrow huge loans from banks.
And Applied Digital isn't the only company to have had success raising capital, either. Just in July this year, another AI data center company, CyrusOne, received a loan of up to US$7.9 billion to support their new projects, and even received US$1.8 billion in capital before that. Such an amount is simply jaw-dropping! CoreWeave, a giant in the GPU rental market, successfully raised US$1.1 billion as early as May, and not long after, it raised another US$7.5 billion in debt from investors.
Of course, in addition to these giants, there are many emerging startups on the market who are also desperately competing for this piece of cake. Foundry, an AI cloud startup, received $80 million in Series A financing before its official launch in August. Groq, a unique chip manufacturer, also successfully raised US$640 million in funding last month with its customized language processing unit (LPU).
To be honest, this is a very hot market today, and almost every data center operator wants to get on the AI train. New companies like TensorWave and Voltage Park are working hard to expand computing clusters based on AMD and Nvidia GPUs, hoping to find their place in this fiercely competitive market.
It is worth mentioning that the reason why Nvidia is willing to invest in these projects is also very obvious. After all, GPU sales are limited by data center capacity. Each time these accelerators are deployed, they generate $1 per hour in subscription revenue for Nvidia. If you have a cluster of 20,000 GPUs, that's not a small amount, and the income is staggering.
All in all, the arrival of the AI era has promoted an unprecedented prosperity in data center construction. Huge investments and fierce competition have combined to create a unique scene in this market. The active layout of companies such as Nvidia also indicates that the AI industry will continue to develop rapidly in the future.