OpenAI gears up to combat AI chip shortages by exploring in-house production

OpenAI, the company behind ChatGPT, is exploring the possibility of producing its artificial intelligence (AI) chips to address the persistent shortage it has been grappling with, Reuters reported on Oct. 6.

OpenAI has reportedly conducted comprehensive market assessments, including possibly acquiring an undisclosed chip manufacturer. However, this plan is still in the process of being finalized, with internal discussions ongoing.

Last year, OpenAI CEO Sam Altman highlighted the importance of acquiring more chips for its operations as the shortages impeded the company’s short-term objectives.

A now-deleted blog post from Raza Habib, the CEO of AI firm Humanloop, said:

“A common theme that came up throughout the discussion was that currently OpenAI is extremely GPU-limited and this is delaying a lot of their short-term plans. The biggest customer complaint was about the reliability and speed of the API. Sam acknowledged their concern and explained that most of the issue was a result of GPU shortages.

AI chips have become a hot commodity since ChatGPT launched to much acclaim. These chips are designed to make building AI systems faster and cheaper. However, with the proliferation of several AI startups, it has become much more difficult for OpenAI to acquire these chips for its chatbot.

Due to this, OpenAI began considering solutions to combat the scarcity issues it faces. Some solutions the company is reportedly considering include the development of proprietary chips, strengthening its partnerships with Nvidia, which dominates more than 70% of the market, and exploring alternative chip suppliers.

Meanwhile, if OpenAI were to begin manufacturing its chip, it would join a small band of tech firms, like Google and Amazon, that produce their chip themselves.

OpenAI has yet to respond to CryptoSlate’s request for comment as of press time.

The post OpenAI gears up to combat AI chip shortages by exploring in-house production appeared first on CryptoSlate.

Credit: Source link