OpenAI is expecting a significant shift in its computing power sources over the next five years, with forecasts indicating that by 2030, approximately 75% of its data center capacity will be supported by SoftBank, a newer financial partner. This transition marks a noticeable departure from its current dependence on Microsoft, which currently handles most of OpenAI's computing needs. While the shift occurs, OpenAI will continue to increase its investment in Microsoft’s data centers and projects significant cost increases, predicting a cash burn of $20 billion by 2027, with operational expenses for AI models eventually exceeding training costs.
OpenAI is shifting its data center reliance from Microsoft to SoftBank, expecting the latter to provide 75% of its computing power by 2030.
Despite the shift, OpenAI will not stop using Microsoft’s data centers immediately and plans to increase spending on them for a few more years.
OpenAI projects that its overall costs will significantly increase, forecasting a cash burn of $20 billion by 2027, compared to $5 billion in 2024.
By 2030, OpenAI anticipates that the expenses for running AI models will surpass the costs of training them, reflecting the growing demand for AI computing.
Collection
[
|
...
]