OpenAI, the renowned artificial intelligence research organization responsible for ChatGPT, is actively exploring the possibility of developing its own AI chips to address concerns over a shortage of essential hardware components. According to sources familiar with the company’s internal discussions, OpenAI is even evaluating potential acquisition targets in pursuit of this ambitious goal.
As of now, OpenAI has not made a final decision to proceed with these plans, but the organization has been engaged in discussions dating back at least a year to mitigate the scarcity of expensive AI chips on which it heavily relies. Options under consideration include designing and manufacturing its own AI chips, fostering closer collaborations with established chipmakers like Nvidia, and diversifying its supplier base beyond Nvidia.
When contacted for comment, OpenAI declined to provide any official statement on these developments.
OpenAI’s CEO, Sam Altman, has publicly emphasized the critical importance of acquiring additional AI chips to sustain the organization’s growth and innovation. He has expressed frustration over the scarcity of graphics processing units (GPUs), which are essential for running AI applications. Currently, Nvidia holds a dominant position, controlling more than 80% of the global GPU market.
This drive to secure more chips is driven by two primary concerns articulated by Altman. First, there’s a significant shortage of advanced processors required to power OpenAI’s software. Second, the costs associated with maintaining the hardware infrastructure necessary for OpenAI’s operations are described as “eye-watering.”
Since 2020, OpenAI has been developing its generative AI technologies on a massive supercomputer constructed in partnership with Microsoft, one of its key supporters. This supercomputer utilizes a staggering 10,000 Nvidia GPUs. The operational expenses for running ChatGPT are substantial, estimated to be around 4 cents per query, according to an analysis conducted by Bernstein analyst Stacy Rasgon. If ChatGPT queries were to scale to just a tenth of Google’s search volume, it would initially require approximately $48.1 billion worth of GPUs and an ongoing annual expense of about $16 billion in chips.
Should OpenAI decide to embark on creating its own custom AI chips, it would join a select group of major tech players, including Alphabet’s Google and Amazon.com, who have pursued chip development to align with their business needs.
However, the prospect of OpenAI developing its own AI chips is a complex and costly undertaking. Industry experts suggest that it could potentially involve annual expenditures in the hundreds of millions of dollars. Furthermore, even with significant resource allocation, success is not guaranteed.
While OpenAI has contemplated the idea of building its custom chips to the extent of performing due diligence on potential acquisition targets, the identity of these targets remains undisclosed.
Even if OpenAI decides to proceed with this initiative, it is expected to be a multi-year endeavor, leaving the organization reliant on commercial chip providers such as Nvidia and Advanced Micro Devices (AMD) in the interim.
It’s worth noting that some major tech companies that have ventured into developing their processors have faced challenges. Meta’s custom chip efforts, for example, have encountered difficulties, leading to the abandonment of certain AI chips. Nevertheless, Meta is now working on a new chip designed to accommodate a wide range of AI tasks.
Interestingly, OpenAI’s primary supporter, Microsoft, is reportedly developing its own custom AI chip, which OpenAI is currently testing. These developments could potentially signify a shift in the relationship between the two organizations.
The demand for specialized AI chips has skyrocketed since the introduction of ChatGPT. These specialized AI accelerators are crucial for training and running the latest generative AI technologies. Currently, Nvidia stands out as one of the few chipmakers producing AI chips, enjoying a dominant position in the market.
As OpenAI continues to explore its options, the potential development of its AI chips remains an exciting development in the rapidly evolving landscape of artificial intelligence technology.