According to a report by The Information websiteOpenAI has recently invested up to US$160 billion (approximately NT$5 trillion) in renting cloud computing resources, which is equivalent to burning tens of millions of dollars every day just to ensure the stable operation of services such as ChatGPT and GPT-5.
Even more astonishingly, OpenAI plans to spend an additional $1000 billion over the next five years specifically on leasing "backup servers" to ensure sufficient computing power. This expense is not even included in the original $3500 billion computing power investment plan.
Computing power investment reaches the highest level in the technology industry
OpenAI CFO Sarah Friar revealed that over the past year, the company has repeatedly delayed the release of new features and models due to insufficient computing power, even having to slow down some product operations to ensure the availability of core services. According to recent plans, OpenAI will invest $3500 billion in cloud server leasing between 2024 and 2030, with annual spending expected to exceed $1000 billion in 2030 alone.
In terms of revenue share, OpenAI's R&D investment is close to 50% of its revenue, far higher than the 10%-20% of Amazon, Microsoft, Oracle, and Alphabet, and even nearly double that of Meta (about 25%), which is famous for its development funding.
Taking this year as an example, of OpenAI's $16 billion expenditure, training costs accounted for $9 billion and inference costs reached $7 billion. It is estimated that by 2030, both investments will be close to $50 billion, about six times that of the GPT-5 era.
Preview new products, Pro subscribers enjoy them first
OpenAI CEO Sam Altman recently announced the release of new "computationally intensive" products in the coming weeks, showcasing the technological breakthroughs that can be achieved through massive computing power. However, speculation has arisen that these new features may be exclusive to ChatGPT Pro members, with some even requiring additional fees, potentially sparking dissatisfaction among some Plus subscribers.
As the scale of investment continues to expand, OpenAI obviously needs to strike a balance between R&D and revenue, accelerating the evolution of AI models while looking for new business models.
