Amidst an industry-wide shortage of GPUs, OpenAI is also mulling making its own chips.

OpenAI is actively considering the development of its own artificial intelligence (AI) chips and has even explored potential acquisition targets in this regard, according to Reuters. While no final decision has been made, the company has been addressing the challenge of the high cost and limited availability of AI chips, which are crucial for its operations. CEO Sam Altman has expressed concerns about the scarcity and cost of graphics processing units (GPUs), which are essential for AI applications. OpenAI currently relies heavily on Nvidia (NASDAQ: NVDA) for these chips, a company that controls a significant share of the global market for AI-capable GPUs.

To tackle this issue, OpenAI is exploring multiple options, including manufacturing its own AI chips and collaborating more closely with chipmakers like Nvidia. Altman views the acquisition of more AI chips as a top priority, considering the steep costs associated with running the hardware necessary for the company’s AI efforts. Currently, OpenAI utilizes a supercomputer powered by 10,000 Nvidia GPUs for its generative AI technologies, including ChatGPT.

Running ChatGPT is expensive, with each query costing around 4 cents. If its queries were to scale to a tenth of Google search, it would require substantial investments in GPUs and chips. Developing custom AI chips would place OpenAI among a select group of tech giants, such as Google (NASDAQ: GOOGL) and Amazon (NASDAQ: AMZN), who design their own chips. However, this endeavor would be a significant strategic initiative involving substantial investments and no guarantee of success. OpenAI is also considering potential acquisitions in this space.

Despite these considerations, OpenAI’s transition to custom chips, whether through development or acquisition, would take several years, leaving the company dependent on commercial providers like Nvidia and Advanced Micro Devices (NASDAQ: AMD) in the interim. Other tech companies have faced challenges in building their own processors, as evidenced by Meta’s struggles with custom chips.

Microsoft (NASDAQ: MSFT), one of OpenAI’s key backers, is also developing a custom AI chip, potentially signaling a divergence in their strategies. The demand for specialized AI chips has surged, particularly with the launch of ChatGPT, which relies on AI accelerators for training and operation. Nvidia remains a dominant player in this field.

Related Story: The Critical Shortage Of AI Chips

Photo by Levart_Photographer on Unsplash