Stargate: $500 billion for AI

What are the potential implications of SoftBank's historic gamble?

Jake Handy·

Image from Stargate: $500 billion for AI

Stargate, a four-year initiative backed by SoftBank, Oracle, OpenAI, and others, was formally revealed yesterday with the intention to “revolutionize AI infrastructure” in the U.S. with a staggering $500 billion investment. First introduced in April of last year (with a $100 billion number attached then), the project promises to build data centers, generate jobs, and push boundaries in artificial intelligence. These hefty promises naturally ignited a debate about its feasibility and potential impact, one that will continue for years.

Here's the core of what you need to know.

The deal

SoftBank leads financing under Masayoshi Son’s chairmanship, with OpenAI driving operational direction. Microsoft, Oracle, NVIDIA, and Arm join as technology partners. An initial $100 billion outlay will expand data-center capacity, develop advanced AI models, and bolster U.S. industrial strength. Key projects include AI model training and infrastructure development, centered initially in Texas.

  • SoftBank. Providing financial backing, despite skepticism about its capacity to raise $500 billion.
  • OpenAI. Overseeing AI development and operational logistics, leveraging partnerships beyond Microsoft to meet escalating compute demands.
  • Oracle. Building data centers, supplementing Microsoft’s cloud dominance.
  • Microsoft. Retaining a central role while ceding some exclusivity (highlighting OpenAI’s need for diversified resources).
  • NVIDIA. Supplying GPUs critical for AI workloads.
  • MGX and Arm. Contributing financial and technical resources.

The risks

  • Funding. Critics doubt the viability of raising $500 billion, noting SoftBank’s mixed history with large-scale investments (hello WeWork).
  • Economic risks. Expensive AI models might not yield proportional returns if generative AI adoption stalls.
  • Overcapacity. Parallel efforts by Microsoft, Google, and Amazon could lead to a glut of underused data centers.
  • Data scarcity. AI research faces a shortage of high-quality data, raising concerns about reliance on synthetic inputs and risks like “model collapse.”