Jensen Huang Touts AGI as Nvidia Moves to Tighten Its Grip on AI
Summary
- The article said the arrival of the AGI era and the unveiling of GPT-6 Astra strengthened the case for expanding demand for Nvidia GPUs.
- It said Big Tech estimates AGI would require at least 1 million to 10 million GPUs, bolstering the case for continued AI infrastructure investment.
- The article said Jensen Huang highlighted the use of more than 100,000 Nvidia Grace Blackwell NVL72 chips and plans to bring an additional 400,000 GPUs online.
Forecast Trend Report by Period
Running AGI Will Require Millions of GPUs
Move aims to calm bubble fears and secure future demand
“The era of artificial general intelligence has arrived. Congratulations, OpenAI team.”
Nvidia Chief Executive Officer Jensen Huang was among the first to congratulate OpenAI after it unveiled GPT-6 Astra on September 3. In a post on X, formerly Twitter, Huang wrote that it had taken four years to go “from ChatGPT to Astra.”
Nvidia invested $30 billion in OpenAI last year and earlier this year. Huang may have publicly congratulated the company with that investment in mind. But many analysts say the move was about more than that, given that the investment is small relative to Nvidia’s $5.26 trillion market capitalization.
There is another motive. Microsoft holds a 27% stake in OpenAI. When Microsoft made its initial investment, the two companies included a clause in their contract stating that Microsoft would lose access to OpenAI technology if OpenAI achieved AGI. The clause was removed in April, but OpenAI had long pursued AGI with that provision as a backdrop. The article said sentiment inside OpenAI had been widespread that reaching AGI would give the company more freedom.
That gives Huang both a rationale and an opportunity to pull a potentially more independent OpenAI deeper into Nvidia’s AI ecosystem. Huang emphasized on X that more than 100,000 Nvidia Grace Blackwell NVL72 chips were used to train Astra.
For Huang, the arrival of AGI could also help quiet increasingly frequent claims that AI is in a bubble. Whenever those concerns grow, investment in AI infrastructure such as data centers tends to weaken. That in turn can diminish the role of Nvidia’s graphics processing units, or GPUs, at the heart of those facilities.
The arrival of AGI stands to drive another massive wave of GPU demand. If generative AI requires 20,000 to 100,000 GPUs, Big Tech estimates AGI would need at least 1 million to 10 million GPUs to operate. That helps explain why Huang said immediately after Astra’s unveiling that another 400,000 GPUs would soon come online.
Ra Hyun-jin, Hankyung.com reporter raraland@hankyung.com
Korea Economic Daily
hankyung@bloomingbit.ioThe Korea Economic Daily Global is a digital media where latest news on Korean companies, industries, and financial markets.