In response to the global scarcity of advanced AI chips, OpenAI, the trailblazing company behind ChatGPT, is actively exploring the possibility of manufacturing its own artificial intelligence chips. Insider sources reveal that the company has not only considered building its own chips but has also evaluated potential acquisition targets, marking a significant shift in its strategy to address the expensive and limited AI chip supply chain.
In recent internal discussions disclosed to Reuters, OpenAI has grappled with the challenge of securing an adequate supply of costly AI chips. CEO Sam Altman, recognizing the critical role of advanced processors in powering OpenAI’s software, has made the acquisition of more AI chips a top priority for the company. The scarcity of graphics processing units (GPUs), particularly dominated by Nvidia with an 80% global market share, has become a major impediment for OpenAI’s operations.
OpenAI’s quest for more AI chips
The shortage of AI chips has prompted OpenAI to explore various options, ranging from building its own chips to collaborating more closely with existing chipmakers like Nvidia. Diversifying suppliers beyond Nvidia has also been under consideration as part of the company’s strategic plan. But, as of now, OpenAI has not made a final decision to move forward with any particular option.
The costs associated with running the hardware necessary to power OpenAI’s ambitious projects, especially the operation of ChatGPT, have been described by Altman as “eye-watering.” Running the language model, developed on a supercomputer constructed by Microsoft and powered by Nvidia’s GPUs, is an expensive endeavor for OpenAI. Analysts estimate that each ChatGPT query costs approximately 4 cents, posing a significant financial burden.
Custom chips era
If OpenAI proceeds with the ambitious plan to develop its own AI chips, it would join the ranks of tech giants like Google and Amazon who have opted to control the design of chips fundamental to their businesses. This strategic move aligns with Altman’s vision to overcome the challenges posed by the scarcity of advanced processors and the soaring costs associated with AI hardware.
Despite the potential benefits of in-house chip production, it remains uncertain whether OpenAI will fully commit to this major initiative. Industry veterans caution that building custom chips is a substantial investment that could amount to hundreds of millions of dollars annually. The path to success is riddled with uncertainties, and even if OpenAI were to allocate significant resources, success is not guaranteed.
Considering the challenges of developing custom chips, OpenAI has also explored the option of acquiring an existing chip company. An acquisition could expedite the process of building OpenAI’s own chip, mirroring the strategy employed by Amazon in its acquisition of Annapurna Labs in 2015. Although details about the potential acquisition target remain undisclosed, it indicates OpenAI’s willingness to explore diverse avenues to secure a stable supply of AI chips.
But, even if OpenAI proceeds with plans for a custom chip, including a potential acquisition, the timeline for implementation is likely to span several years. This leaves the company dependent on commercial providers like Nvidia and Advanced Micro Devices in the interim. Some major tech players have encountered challenges in developing their own processors, as seen in Meta’s custom chip efforts, which faced setbacks leading to the abandonment of certain AI chips.
Changing dynamics with Microsoft and industry demand
Microsoft, OpenAI’s primary backer, is also actively involved in the development of custom AI chips that OpenAI is currently testing. This development could signal a shift in the dynamics between the two companies, potentially leading to further distancing as OpenAI explores self-sufficiency in AI chip production.
The demand for specialized AI chips has surged since the launch of ChatGPT, necessitating the use of AI accelerators for training and running the latest generative AI technology. Nvidia, as one of the few chipmakers producing effective AI chips, currently dominates the market. OpenAI’s strategic considerations reflect the evolving landscape of AI technology and the company’s determination to secure a reliable and cost-effective supply of AI chips to sustain its groundbreaking projects in the long run.
