The center, set to become the ChatGPT maker’s largest, is being constructed in the Texas community of Abilene by the startup Crusoe and is critical to boosting OpenAI’s computing power, the Wall Street Journal (WSJ) reported Wednesday (May 21).
Crusoe told the news outlet that the funding, a mix of debt and equity, will expand the data center to eight buildings from two and bring the amount secured for the project to $15 billion.
The data center is slated for completion next year. Each building will run up to 50,000 Nvidia Blackwell chips, typically used for training large language models.
As the WSJ noted, the development of this site is critical to reducing OpenAI’s dependence on Microsoft in the race to build the next generation of artificial intelligence (AI) models.
Eventually, the report added, OpenAI hopes to construct and operate its own data centers. Earlier this year, the company teamed with Japanese tech investor SoftBank and cloud provider Oracle on “Stargate,” an AI infrastructure project with a $500 billion price tag.
In other AI news, PYMNTS wrote earlier this week on new research that shows the limits of AI agents. This study evaluated how these agents performed in a simulated work environment and found that even the most advanced systems can only complete a handful of tasks.
“They get stumped by tasks that are easy for humans, such as closing a pop-up window or waiting 10 minutes before escalating an issue,” that report said.
Researchers from Carnegie Mellon University and other institutions created a software company staffed by 17 AI agents who interacted with AI agents from popular AI models such as OpenAI’s GPT-4o and Anthropic’s Claude 3.5 Sonnet.
The researchers then evaluated how well the AI models performed on tasks in fields like software engineering, project management, finance and HR.
The conclusion? While AI agents can improve the productivity of workers, “they’re not ready to replace real-world human jobs,” study co-author Boxuan Li told PYMNTS.
The results stand in contrast to the vast amounts of money being invested in AI agents by companies looking to tap into the promise of smart automation.
“So much money is going to the agent area,” noted co-author Yufan Song in an interview with PYMNTS. Agents “can help speed up our productivity, but to replace humans, I think, still needs some time.”
