A brand new report from Bloomberg says that after once more OpenAI CEO Sam Altman’s efforts to boost billions for an AI chip firm are geared toward utilizing that cash to develop a “community of factories” for manufacturing that might develop throughout the globe and contain working with unnamed “prime chipmakers”. .”
A significant value and limitation to operating AI fashions is having sufficient chips to deal with the computations behind bots like ChatGPT or DALL-E that reply to requests and generate photographs. Nvidia’s worth rose above $1 trillion for the primary time final yr, partly due to a digital monopoly it has because the GPT-Four, Gemini, Llama 2 and different fashions rely closely on its fashionable H100 GPUs.
Consequently, the race to fabricate extra high-powered chips to run advanced AI programs has intensified. The restricted variety of factories able to producing high-end chips causes Altman or anybody else to bid for capability years earlier than they want it to make the brand new chips. And going up in opposition to the likes of Apple requires deep-pocketed traders to bear prices that OpenAI cannot but afford. SoftBank Group and Abu Dhabi’s AI holding firm G42 have been in talks to boost cash for Altman’s mission.
AWS, Azure and Google additionally use Nvidia H100 processors. Meta CEO Mark Zuckerberg stated this week The Verge reporter Alex Heath that “by the top of this yr, Meta could have greater than 340,000 Nvidia H100 GPUs” as the corporate pursues the event of synthetic basic intelligence (AGI).