Breaking News: OpenAI Is Jumping on the Trend of Developing Its Own Custom AI Chip

Following reports by Reuters on Oct 5, OpenAI, the company behind the artificial intelligence chatbot ChatGPT, claims to start exploring the possibility of producing processing chips in-house amid a global shortage of expensive and difficult-to-produce hardware.

According to a recent internal discussion reported by Reuters, OpenAI has yet to give the green light to advance the project but has proposed various options to solve the shortage of expensive AI chips the company heavily relies on. For example, the proposed ideas include developing its own AI chip, working more closely with other chipmakers, including NVIDIA and diversifying its suppliers beyond the leader of the AI Chips.

OpenAI CEO Sam Altman Places Priority in Acquiring More AI Chips

To fund this supposed project, OpenAI CEO Sam Altman has made acquiring more AI chips a top priority for the company. According to Altman, he identifies two main concerns in the race to get more chips. 

• A shortage of advanced processors powering OpenAI’s software
• The exorbitant costs associated with running the hardware needed to power its efforts and products.

Unsurprisingly, OpenAI requires a large sum of money to run its company, as it uses around 10,000 graphics processing units from NVIDIA to utilise its generative artificial intelligence technologies on a massive supercomputer constructed by Microsoft. Based on analysis by Bernstein analyst Stacy Rasgon, if ChatGPT queries grew to 1/10 the size of Google Search, it would initially require about $48.1 billion in GPUs and about $16 billion in chips per year to maintain operations.

In an off-the-record meeting in March, OpenAI CEO Sam Altman allegedly claimed that a computer chip shortage is hindering the progress of ChatGPT. In a now take-down post of Raza Habib, the OpenAI CEO continuously uttered that OpenAI is extremely GPU-limited and is delaying many of their short-term plans in the discussion.

Related: Why China Remains Hungry for AI Chips Despite U.S. Restrictions

OpenAI Joins Google and Amazon in Exploring Custom Chips Era

It is much better late than ever if OpenAI decides to join massive tech players such as Google and Amazon to initiate designing the chips that are fundamental to their businesses. However, it is not a guarantee of success, as it will take a lot of money, resources and time to develop an AI chip that does not dull to NVIDIA. For example, Meta struggled with issues when building the chip from scratch.

Furthermore, the main backer of OpenAI, Microsoft, has recently developed an AI Chip, which OpenAI is testing. It could mean there is minimal chance of the two companies supporting one another. Even worse, it could mean a further distancing between the two.

On the other hand, if they follow in the footsteps of Amazon acquiring a chip company, which they did back in 2015, it will massively quicken the process and reduce the likelihood of the project backfiring.

Related: Amazon Commits $4 Bn Investment in Anthropic, an OpenAI-Rival: 5 Important Points

Concluding Statement

After the successful launch of ChatGPT last year, the demand for specialised AI chips has soared. If you know economics, you will understand the effect of supply and demand. In the latest generative AI technology, specific AI chips are a necessity. As NVIDIA dominates the market with its AI chips, OpenAI is now on the back foot, mainly due to its own success. What do you think? Let us know your thoughts on our Facebook, Instagram and Twitter pages.

Frequently Asked Questions

Who Is the Leader in AI Chips?

NVIDIA is the simple and correct answer. Although Google, Amazon, Meta, and others have also produced AI chips, NVIDIA accounts for more than 70 per cent of the sales and leads the position in training generative AI models.

Why Is AI Chip Important?

An AI Chip is a thousand times as efficient as a CPU as it provides power efficiency and high performance for AI applications because of their heterogeneous computing abilities. Thus, many generative AIs, such as Gemini and ChatGPT, require AI chips to function and run.

What Challenges Does OpenAI Face in Developing Their Own AI Chip?

As mentioned above, developing an AI chip from scratch requires high cost and time with an uncertain outcome. Furthermore, a failure here will result in competitors overtaking them, and it is not something they want to happen as ChatGPT is still the leading LLM.