“Massive $100 Billion AI Infrastructure Initiative Set to Outpace Chinese Competitors”

Oracle, OpenAI, along with investors from Japan and the UAE, have initiated a $100 billion project to construct data centers for the operation of AI applications. This signifies a shift in the U.S.-China contest for artificial intelligence towards raw computing capability over sophisticated programming.

As part of the initiative named Stargate, Oracle has commenced the development of ten data centers in Texas, each covering half a million square feet, CEO Larry Ellison announced on Tuesday during a White House press briefing.

“That number will rise to 20, along with additional sites,” Ellison stated.

This initiative is being managed by OpenAI, a generative AI firm, which will primarily utilize these data centers, according to a source familiar with the arrangement who spoke to Defense One.

“I believe this will be the most significant project of this period,” said OpenAI’s founder Sam Altman at the press conference. “For [artificial general intelligence] to develop here, this will generate hundreds of thousands of jobs and foster a new industry located here.”

The principal backers are private enterprises, among them the Japanese holding conglomerate SoftBank and MGX based in the United Arab Emirates. Their initial contribution will amount to $100 billion, and Stargate aims to attract $500 billion within five years.

The press briefing also showcased President Donald Trump, who claimed responsibility for organizing the venture but did not specify what role the U.S. government might play in Stargate. 

“What we aim to do is to retain this within our nation. China represents competition,” Trump noted.

More about essential matters, less about numbers

Officials from varied political backgrounds in the U.S. have stressed the necessity of maintaining the U.S. lead in AI over China. Last week, Jake Sullivan, the outgoing national security advisor in the Biden administration, forecasted “disaster” if the U.S. fails to hold its market and research superiority along with leading the establishment of global standards regarding AI usage.

In an interview with Axios, Sullivan indicated that the repercussions would be “dramatic, and extremely detrimental—inclusive of the democratization of very potent and lethal weaponry; severe disruption and dislocation of employment; an overwhelming influx of misinformation.”

In 2017, China unveiled a strategy to emerge as a “global leader” in AI by 2050. Five years later, the Special Competitive Studies Project reported that the U.S. maintained a slight edge in AI research. China has since intensified its efforts; last August, the Chinese government announced an investment of 43.5 billion yuan (US$6.1 billion) for the establishment of eight new computing centers. 

On Tuesday, SCSP President Ylli Bajraktari informed Defense One, “While the U.S. has distinct advantages, China is quickly narrowing the gap. Specifically, China has shown proficiency in swiftly executing extensive infrastructure initiatives. Thus, the specifics of this announcement will be pivotal, yet it’s a crucial strategic advancement in the appropriate direction.”

Bajraktari remarked that the competition in AI is evolving, shifting from the intricacies of mathematical models towards the tangible physics of energy and data storage. 

“The principal obstruction to AI progress in the United States is not an absence of innovation, but rather a deficiency in infrastructure. This encompasses not only data centers but also the energy grids and high-speed networks imperative for their support,” he stated.

Greg Allen, who leads the Wadhwani AI Center at the Center for Strategic and International Studies, echoed this sentiment. 

“The focus of strategic rivalry in AI is transitioning away from software engineering secrets that nations attempt to conceal, towards large-scale hardware AI infrastructure that nations endeavor to construct and implement,” Allen remarked. Daniel Castro, vice president at the Information Technology and Innovation Foundation, expressed that the White House announcement is an encouraging indication.

“This collaboration signifies a mutual objective: establishing the groundwork for the U.S. to flourish in the AI economy. Although the global AI competition won’t be concluded overnight, achieving significant milestones on the first day of this new administration illustrates that the U.S. tech sector is revitalized and prepared to take decisive measures forward with this new administration,” Castro stated.

Nevertheless, the Biden administration also made considerable efforts to bolster the U.S. position in AI. It initiated a $30 million National AI Research Resource pilot project last March, proposed a $3 billion increase in federal AI usage, and implemented export controls to hinder China’s access to advanced AI-relevant computer chips. Additionally, the Biden White House released an executive order focused on AI risks, alongside a national security memorandum addressing AI.

On Monday, Trump revoked the executive order; however, Allen from CSIS suggested that it shouldn’t be overly emphasized.

“Much of the directed federal action resulting from the Biden administration’s AI executive order has already been executed in a timely manner,” he noted, adding: “It’s worth noting that President Trump did not revoke the AI national-security memorandum.”