SoftBank Group Corp. is exploring project financing, a technique often used in large-scale infrastructure projects, to fund its ambitious $500 billion artificial intelligence venture with OpenAI, known as Stargate Project. This strategy could significantly reduce upfront investment requirements for SoftBank and its partners.
Project financing allows for long-term funding based on projected cash flows, similar to how the Trans-Alaskan pipeline was financed. SoftBank is considering incorporating aspects of this structure into Stargate, which involves multiple datacenter and power generation projects. One potential scenario involves SoftBank, OpenAI, Oracle Corp., and Abu Dhabi’s MGX contributing approximately 10% of the total cost in equity, with the remaining funds sourced from debt markets.
This ambitious project, announced last month, aims to build tens of gigawatts of datacenter capacity, potentially the largest computing power rollout in history. Stargate has already broken ground on its first datacenter in Abilene, Texas, and is evaluating additional sites in over a dozen other states. Initial spending is projected at $100 billion.
However, the funding specifics remain largely undetermined. SoftBank is actively seeking additional stakeholders and exploring various funding options. These discussions come amidst a changing AI landscape, with the rise of DeepSeek’s low-cost, open-source AI posing a competitive challenge. Furthermore, an unsolicited bid for OpenAI led by Elon Musk adds further complexity to Stargate’s fundraising efforts.
A combined 10% equity stake in Stargate would represent roughly $50 billion. Based on potential equity splits, SoftBank and OpenAI’s individual contributions could range from $15 billion to just under $25 billion each, assuming smaller contributions from Oracle and MGX. The final funding structure and the impact of emerging market dynamics on the project’s long-term viability remain to be seen.
