OpenAI expects to burn $115 billion on ChatGPT through 2029 — but is developing its own AI chips to help offset soaring costs

by

As artificial intelligence technology keeps progressing and reaching impressive milestones, such as with OpenAI, we’re witnessing that joining this sector can be expensive. A recent report by The Information predicts that the company behind ChatGPT will likely expend approximately $115 billion by 2029 while intensifying its AI endeavors.

It’s troubling to note that the report suggests the projected cost of $115 billion is $80 billion greater than the company’s prior estimation. To give you some perspective, the AI company has already spent around $8 billion in 2022, which is about $1.5 billion more than the amount they had initially forecasted for this year.

It appears that OpenAI is taking steps to address any potential problems, such as creating their own server chips and data centers for their future technology, among other strategies they’ve outlined.

Next year, the company behind ChatGPT is teaming up with Broadcom, a leading American semiconductor corporation, to unveil their first AI-integrated chip. Yet, according to The Financial Times, OpenAI intends to utilize these chips internally and won’t offer them as a product for customers to purchase.

This year, it’s been rumored that Microsoft backed out of two large data center agreements due to its decision not to offer further assistance in training ChatGPT. Yet, OpenAI CEO Sam Altman seemed to qualify these statements, suggesting that the company may no longer be limited by computational resources.

In July, a partnership was formed between OpenAI and Oracle, aiming to expand Stargate data center capacity by 4.5 gigawatts within the U.S. Additionally, OpenAI appears to be considering Google for their cloud computing needs as they intensify their efforts in artificial intelligence.

After OpenAI revealed their massive $500 billion initiative, the Stargate project, aimed at building data centers nationwide in the U.S. for powering advanced technologies, this development led to a change in Microsoft’s previous exclusive role as OpenAI’s cloud provider. However, Microsoft still retains the option to reject any future collaborations.

According to Salesforce CEO Marc Benioff, it’s expected that Microsoft will diminish their reliance on OpenAI’s technology in the future. Already, the tech titan appears to be taking steps towards a more balanced approach, rather than relying excessively on OpenAI.

Mustafa Suleyman, the CEO of Microsoft’s AI division, shared that the company is working on advanced AI models, but they may be about 3 to 6 months behind those developed by OpenAI. Suleyman suggested that Microsoft aims to take a strategic second position in the field while keeping costs reasonable. Additionally, Microsoft is experimenting with third-party AI models within Copilot, which could help the company reduce its dependence on OpenAI.

OpenAI has a big money problem

It’s worth noting that OpenAI recently secured a significant $40 billion investment, primarily funded by SoftBank. This substantial fundraising drive has boosted OpenAI’s market value to an impressive $300 billion. However, this new financial standing puts the AI company in a challenging position, given the mounting pressure from investors to transform into a revenue-generating entity before the end of this year.

If the company doesn’t reach this critical point, it may face potential issues such as loss of investor support, interference from external parties, and hostile takeovers. According to industry experts and financial analysts, investor interest in AI is declining rapidly, suggesting they might soon redirect their funds to other ventures. This could potentially lead to Microsoft acquiring OpenAI within the next three years.

Despite Microsoft appearing to limit its support, OpenAI seems to be hindered in transforming into a profitable organization. This situation has led to growing tension within their multi-billion dollar partnership, which may strain what was previously a strong tech alliance.

A different document stated that OpenAI might take legal action due to Microsoft’s allegedly anticompetitive business tactics. On the contrary, Microsoft has signaled its readiness to completely terminate talks with OpenAI and continue the partnership until 2030 as is.

If OpenAI manages to achieve artificial general intelligence (AGI) before 2030, it has the option to break ties with Microsoft earlier. Alternatively, some speculate that OpenAI might claim AGI prematurely by releasing an AI coding agent that surpasses the abilities of a highly skilled human programmer.

The aim was to maintain the original meaning while making the sentence more natural and easy to understand.

On previous occasions, OpenAI has encountered similar predicaments. Just last year, rumors surfaced suggesting financial instability, with estimates indicating potential losses of around $5 billion within a year. To put this into perspective, the company’s annual expenses, which include model training costs amounting to $7 billion and a staff budget of $1.5 billion, significantly outweigh its $3.5 billion in earnings.

Read More

2025-09-09 14:40