Behind $500 billion AI data center plan, US startups jockey with tech giants 

Published 01/23/2025, 01:44 PM
Updated 01/23/2025, 01:51 PM
© Reuters. FILE PHOTO: AI Artificial intelligence words, miniature of robot and U.S. flag are seen in this illustration taken December 21, 2023. REUTERS/Dado Ruvic/Illustration/File Photo
ORCL
-
NVDA
-
EQIX
-

By Stephen Nellis and Anna Tong

SAN FRANCISCO (Reuters) - - This week's announcement by President Donald Trump of a massive private-sector investment to build more AI data centers casts a spotlight on a relatively small and nimble class of cloud computing firms positioned to play a bigger role in the tech sector.

On Tuesday, Trump said that ChatGPT creator OpenAI, SoftBank (TYO:9984), Oracle  and others will pour up to $500 billion in private capital into a joint venture called Stargate, which he said will build data centers and create more than 100,000 jobs in the United States. One goal of the project, said Trump, is to keep the U.S. ahead of China in the global race for supremacy in artificial intelligence.

Missing from the announcement was San Francisco-based startup Crusoe, which was tapped by Oracle (NYSE:ORCL) to build the first data center for Stargate, two sources familiar with the matter told Reuters. In order to move more quickly and keep capital expenditures down, Oracle contracted out its development to Crusoe and then engaged OpenAI in talks to be a customer, one source said.

Oracle did not return a request for comment. Tech news website The Information last year reported that Oracle had contracted Crusoe to build a data center in Abilene, Texas, for OpenAI.

Crusoe is among a crop of newer companies - which according to chip research firm SemiAnalysis also includes CoreWeave, Nebius Group and Lambda - that are building cloud computing offerings specifically for the needs of AI companies. They accomplish this usually by amassing huge troves of Nvidia (NASDAQ:NVDA) chips linked for specific kinds of AI work.

In Silicon Valley, these firms are known as "neoclouds" because their AI focus sets them apart from such cloud giants as Microsoft (NASDAQ:MSFT) , Alphabet (NASDAQ:GOOGL)'s Google and Amazon (NASDAQ:AMZN) Web Services, all of which serve both general corporate customers and their own parent companies.

"This is a potential wakeup call that smaller companies can move faster," said Robert Brooks IV, vice president of revenue at Lambda, referring to the Stargate plan. "When AWS and (Google Cloud Platform) build the data center, they think about all these extracurriculars unrelated to AI." Lambda has not disclosed any involvement in the Stargate project. 

The neocloud companies are also routes to market for Nvidia to sell chips to developers. Microsoft, AWS and Google all offer Nvidia chips but also have their own proprietary AI chips that compete against Nvidia's chips.

Oracle and the neoclouds, by contrast, have tended to work closely with Nvidia rather than offer their own alternative chips.

"It feels like (Nvidia) got another big purchaser," said Stacy Rasgon, an analyst at Bernstein who follows the semiconductor industry. "These are not players that look like they're building their own chips at this point."

MOVING DIRT TO MOVE FAST 

To be sure, the neocloud companies themselves may need help moving fast. While Crusoe is building its own Abilene data center that will hold 100,000 specialized AI chips per building when it comes online this year, many of the other neoclouds have turned to leasing existing space or tapping data center construction specialists to help build their sites.

At least in the short term, many benefits may accrue to some existing data center construction firms such as Equinix (NASDAQ:EQIX) or Compass.

"The scale that we're talking about here has never been done before, so it's a huge undertaking that will absolutely create a bunch of jobs and great things like that," said Jason Hardy, chief technology officer for AI at Hitachi (OTC:HTHIY) Vantara, which provides storage systems for data centers. "But I think it's too early to actually tell what this will all flesh out to, other than the fact that this isn't a hyperscaler," Hardy said, referring to incumbent cloud computing firms.

Moreover, Elon Musk earlier this week raised questions about the ability of SoftBank and OpenAI to fund the project. Axios later reported that a portion of the funding could be raised as debt, and tech publication The Information reported that each company would put $19 billion toward the effort. 

But neocloud firms are moving fast. Crusoe was able to hand over the first building in the Abilene data center for tenant improvements within six months - something that would have normally taken several years, Crusoe CEO Chase Lochmiller told Reuters earlier this month. 

© Reuters. FILE PHOTO: AI Artificial intelligence words, miniature of robot and U.S. flag are seen in this illustration taken December 21, 2023. REUTERS/Dado Ruvic/Illustration/File Photo

Crusoe is developing pre-fabricated components that speed up data center construction, similar to how pre-fabricated homes reduce on-site construction time, Lochmiller said.

CoreWeave, another neocloud company, is targeting an initial public offering this year at a valuation of $35 billion, Reuters reported in November.

Latest comments

Risk Disclosure: Trading in financial instruments and/or cryptocurrencies involves high risks including the risk of losing some, or all, of your investment amount, and may not be suitable for all investors. Prices of cryptocurrencies are extremely volatile and may be affected by external factors such as financial, regulatory or political events. Trading on margin increases the financial risks.
Before deciding to trade in financial instrument or cryptocurrencies you should be fully informed of the risks and costs associated with trading the financial markets, carefully consider your investment objectives, level of experience, and risk appetite, and seek professional advice where needed.
Fusion Media would like to remind you that the data contained in this website is not necessarily real-time nor accurate. The data and prices on the website are not necessarily provided by any market or exchange, but may be provided by market makers, and so prices may not be accurate and may differ from the actual price at any given market, meaning prices are indicative and not appropriate for trading purposes. Fusion Media and any provider of the data contained in this website will not accept liability for any loss or damage as a result of your trading, or your reliance on the information contained within this website.
It is prohibited to use, store, reproduce, display, modify, transmit or distribute the data contained in this website without the explicit prior written permission of Fusion Media and/or the data provider. All intellectual property rights are reserved by the providers and/or the exchange providing the data contained in this website.
Fusion Media may be compensated by the advertisers that appear on the website, based on your interaction with the advertisements or advertisers.
© 2007-2025 - Fusion Media Limited. All Rights Reserved.