OpenAI aims to develop its own AI chip on top of "iPhone for AI"

The AI development boom has led to chip shortages and soaring prices, bringing opportunities for newcomers.

According to reports, developers of ChatGPT are considering manufacturing their own AI chips due to the shortage of processors required to handle the powerful software developed by OpenAI and the high costs associated with its operation.

As the company behind ChatGPT, OpenAI has relied on chips from NVIDIA since its inception. NVIDIA holds over 80% of the global AI chip market share and enjoys a quasi-monopoly in this field. However, OpenAI CEO Sam Altman recently commented on the scarcity of these chips. Last week, sources revealed to Reuters that Altman has been exploring the possibility of OpenAI entering the chip business or acquiring existing chip manufacturing companies. He is also considering collaborations with companies other than NVIDIA to diversify suppliers.

OpenAI uses Microsoft's supercomputers, which incorporate 10,000 NVIDIA GPUs (graphics processing units). According to Bernstein analyst Stacy Rasgon, each ChatGPT query costs the company about 4 cents, and this amount could become astronomically large as ChatGPT continues to evolve.

Of course, the enormous cost of entering the chip business represents a significant barrier. Building AI chips in-house would require an investment of hundreds of millions of dollars and take several years. In addition, there are reports that OpenAI is in discussions with former Apple designer Jony Ive about creating an "AI version of the iPhone," a ambitious plan that would undoubtedly require billions of dollars in investment.

For OpenAI, acquiring a chip company seems to be a more feasible option for producing its own chips. Both Google and Amazon control chips that power their operations. Amazon achieved this goal in 2015 through the acquisition of Annapurna. Meta has also explored this option and has launched its own chips, but the company is undergoing a redesign due to some technical issues.

Meanwhile, OpenAI will continue to rely on the leading giants in GPU technology.