Nuked 
Hey there, tech enthusiasts! Buckle up, because we’re diving into some fascinating news regarding OpenAI’s computing strategy.
Recently, OpenAI revealed plans to significantly change how it sources its computing power over the next five years. According to a report from The Information, a major transition is on the horizon.
By the year 2030, OpenAI anticipates that a whopping three-quarters of its data center capabilities will be sourced from a project called Stargate. This initiative is expected to get substantial financial backing from SoftBank, who is becoming one of OpenAI’s key investors.
Currently, Microsoft plays a crucial role as OpenAI’s largest shareholder and primary provider of computing power. However, this situation is poised to transform as OpenAI still plans to ramp up its investments in Microsoft’s data centers in the near term.
But here’s the catch – although the shift towards SoftBank is inevitable, it won’t happen at lightning speed. In fact, OpenAI projects its overall expenditure on computing power to rise dramatically, with cash burn increasing to about $20 billion by 2027. That’s quite a leap from the $5 billion reported in 2024.
As we look toward 2030, a notable trend is expected to emerge: OpenAI’s costs related to running AI models will outstrip what they spend on training these models. It’s an exciting yet challenging time for the AI landscape, and we’ll be watching closely as OpenAI navigates these changes!
