Share this article on:
Father of ChatGPT Sam Altman is looking to source a meagre US$5 trillion to US$7 trillion (roughly A$7.7 trillion to A$10.8 trillion) with which to completely change the nature of artificial intelligence (AI) development.
The need for such a large sum comes from AI’s biggest developmental limitation – computing power. AI is dependent on powerful GPU (graphics processing unit) chips for its development, and stocks are in short supply.
The OpenAI chief executive is currently in talks with global investors, including those from the United Arab Emirates (UAE), to source the money, according to The Wall Street Journal.
For context, US$7 trillion is an unbelievable amount of money. The most valuable company in the world, Microsoft, has a market cap of just over US$3 trillion. Even more notably, global chip sales in 2023 reached US$526.8 billion, over 13 times less than what Altman wants.
However, with that money, Altman wants to completely flip the global chip manufacturing industry on its head.
AI developers have presented concerns over the supply of GPUs, and the AI arms race between the US and China has been largely fuelled by access to GPUs.
Altman’s project would turbocharge the world’s GPU building capacity and output, speeding up AI development by providing developers with a surplus of chips rather than having them fight over the dwindling supply as they are now.
we believe the world needs more ai infrastructure--fab capacity, energy, datacenters, etc--than people are currently planning to build.
— Sam Altman (@sama) February 7, 2024
building massive-scale ai infrastructure, and a resilient supply chain, is crucial to economic competitiveness.
openai will try to help!
“OpenAI has had productive discussions about increasing global infrastructure and supply chains for chips, energy and data centres –which are crucial for AI and other industries that rely on them,” said a spokesperson for OpenAI.
“We will continue to keep the US government informed given the importance to national priorities, and look forward to sharing more details at a later date.”
The other piece in Altman’s puzzle is power. While building the infrastructure for an all-powerful and efficient chip manufacturing system makes sense, AI, GPUs, and their manufacturing and development processes use massive amounts of power.
Altman has already made major investments in start-up companies seeking to make cheap and efficient energy through nuclear fusion, according to The Wall Street Journal.
As part of his plan, Altman hopes to partner with chip manufacturers and energy providers, which would collaboratively fund chip factories to be run by existing chip manufacturing giants, according to sources close to the matter. OpenAI would then be a major customer of the partnership, sourcing its GPUs from the new project.
The plan has been criticised by NVIDIA CEO Jensen Huang, who said the US$7 trillion price tag is way too high.
“You can’t assume just that you will buy more computers,” Huang told UAE AI minister Omar Al Olama.
“You have to also assume that the computers are going to become faster, and therefore, the total amount that you need is not as much.”
Huang said the cost of building AI data centres should be roughly US$2 trillion by 2029.
“There’s about a trillion dollars’ worth of installed base of data centres,” he said.
“Over the course of the next four or five years, we’ll have $2 trillion worth of data centres that will be powering software around the world.”
The GPU shortage and AI race have fuelled additional tensions between China and the US as both countries scramble for tech dominance.
Majority of the world’s chips are made in Taiwan, with Taiwan Semiconductor Manufacturing Company (TSMC) accounting for 50 per cent of the world’s chips.
In what is known as the “Silicon Shield”, Taiwan’s role as a chip manufacturing giant has resulted in the US and its allies being majorly interested in keeping the country independent.