SoftBank’s Masayoshi Son Says A.I. Will Soon Be 10,000 Times Smarter Than Humans
ChatGPT-maker OpenAI is flush with cash, thanks to the likes of Masayoshi Son’s SoftBank. The Japanese conglomerate chipped in $500 million during a recent funding round that valued the Sam Altman-led company at a staggering $157 billion. Announced yesterday (Oct. 2) and led by Thrive Capital, the $6.6 billion round also drew in A.I.-focused companies like Microsoft (MSFT) and Nvidia (NVDA) and investment firms such as Tiger Global Management, ARK Invest and Altimeter Capital.
Son didn’t directly address his investment while speaking today (Oct. 3) at the SoftBank World 2024 conference. But he praised OpenAI and the capabilities of the company’s recently released GPT-o1 A.I. models, codenamed “Strawberry,” describing them as having “developed the ability to think.” Son said he recently tested out the reasoning model by asking it a series of difficult questions, such as how to increase his savings account to 100 million Japanese yen ($681,000), and was impressed by its answers. Son, currently the second richest person in Japan, is estimated to be worth $31.4 billion.
A.I. is on track to become even more powerful in the near future, Son said. His vision for A.I. includes personal agents that will be able to conduct a variety of everyday tasks, ranging from grocery shopping and making reservations to aiding in investments and monitoring the health of family members. “This technology will evolve to a point where your happiness will be its greatest reward,” said Son, adding that human happiness will be the “greatest reward” for personal agents.
Such comments aren’t out of character for Son, who is known for his unrelenting optimism when it comes technological developments. This attitude has caused SoftBank’s Vision Funds to make ambitious tech bets that typically realize either major gains or losses. During the dot-com crash, for example, SoftBank lost some $170 billion on its internet investments. Son also infamously shed some $14 billion after backing the now-bankrupt WeWork. While Son spent the past few years laying low following a series of losses, he told investors last year that SoftBank was going on “offense mode” and was preparing to invest heavily in A.I., having amassed more than $35 billion since 2020.
According to Son, A.I.’s developments will come fast and furious. Artificial general intelligence (A.G.I), a type of A.I. that matches the intelligence of humans, will come to fruition in the next two to three years, Son said at today’s event. And artificial superintelligence, (A.S.I.), which is 10,000 times more capable than humans, can be expected to be achieved in a decade, he claimed. Noting that the human brain consists of some 100 trillion synaptic connections, referring to points where neurons communicate with each other, Son said that “generative A.I.’s parameters—which are equivalent to these synapses—are growing at a furious pace.” (GPT-4 reportedly has about 1.7 trillion parameters.)
To take advantage of the technology’s potential, the SoftBank CEO has ambitious plans to turn his conglomerate into an A.I. powerhouse. SoftBank’s A.I. strategy is currently anchored by Arm, a U.K.-based chip designer it acquired for $32 billion in 2016. Son has previously laid out his desire to eventually establish a variety of A.I.-focused divisions within SoftBank, complete with sectors dedicated to A.I. chips, data centers and even industrial robots. Besides SoftBank’s recent backing of OpenAI, the conglomerate in 2024 alone has also invested in Perplexity AI, an A.I.-powered search startup, and acquired Graphcore, a chipmaker and Nvidia rival.