Masayoshi Son, Japan’s Most Notorious Investor, Is Betting Everything on American AI

They may be digs that hark back to a bygone era, but from his lofty perch Son is very much focused on the horizon. And the next one. After making a fortune in software and parlaying that into telecoms and a raft of tech ventures, Son is now pivoting SoftBank’s $180 billion war chest into artificial intelligence (AI) with unbridled zeal. Deals include taking control of chip firms Arm, Graphcore, and Ampere Computing, as well as self-driving car start-up Wayve, $2 billion into Intel, plus some $40 billion into ChatGPT creator OpenAI—with reports of an additional $30 billion for the latter in the pipeline.

“ChatGPT can give me quicker, more precise, more detailed answers,” Son, 68, gushes. “I can debate, bounce ideas, brainstorm, it’s amazing. But this is the beginning. It’s going to be a billion times more [smart]. Super intelligence is coming soon.”

Artificial super intelligence, or ASI, will become “10,000 times smarter” than humans within a decade, Son predicts, grandiosely telling a shareholders' meeting that he was “born to realize” the breakthrough. ASI combined with physical AI such as humanoid robotics will comprise 10% of global GDP in 10 to 15 years, he tells TIME, rising to 30% over 30 years. “AI models are evolving very quickly,” he says. “So this is going to be a massive industry.”

Exciting stuff. Though Son’s high-risk, even higher-leveraged investment style has courted disaster almost as much as success. His $20 million investment in Chinese e-commerce giant Alibaba was worth close to $200 billion at its peak and arguably history’s greatest investment. But the $18.5 billion he plowed into the now bankrupt office-sharing venture WeWork lists among history’s most foolhardy, spurring Son to sequester himself from public view for 18 months.

After a period of intense scrutiny and losses, Son’s mojo has returned. SoftBank shares hit a record high in late October on the back of the AI boom, briefly propelling Son to be once again Japan’s richest man. Last week, Softbank posted a $20.7 billion profit for the nine months through December—about five times what it recorded the previous year. It’s a vindication for Son, who decades before the explosion of gen AI was preaching about “singularity” and prophesying a time when machines dwarf human intellect. Still, his phoenix-like ability to survive and renew will be tested by his latest venture, which is arguably his most ambitious and consequential.

“He was very excited,” Son says of Trump’s embrace of Stargate. “He's really on top of it. Nothing will slow him.”

Few people spot the next big trend as consistently as Son. To his fans, he’s an unbridled genius of unrivaled prescience. To his detractors, he’s an inveterate gambler who has ridden his luck as far as it can take him. The difference today is that Son is not just betting SoftBank’s fortunes, but those of the American economy and people. Top economists warn that over-hyped AI stocks are a bubble that will surely burst and with it render Stargate a half-a-trillion dollar white elephant.

Not that Son agrees. Asked what AI will transform, he replies: “Everything, every industry. Even farming or fishing or fashion or media. Almost all human activities eventually will be some kind of collaboration with superintelligence and physical AI. It’s just a matter of time. Everybody has to wake up. This is not a bubble.”

Son arrives for our interview dressed in a turtleneck and slacks. A stocky man with receding hair, he shuffles rather than strides, and these languorous movements feel incongruous with a reputation for heart-stopping risk and seizing the next great trend. But they lend to the Zen-like mystique surrounding Son, who’s been described by partners as Yoda and compared himself to Napoleon, Genghis Khan, and even Jesus Christ (whom, he once told a stunned earnings call, had been similarly misunderstood).

Son was born the grandchild of immigrants from Korea in a small town on Japan’s southernmost island of Kyushu. His childhood home was a shack on a plot of unregistered land alongside dozens of similarly makeshift homes. As Korean-Japanese, his family faced discrimination and lived under an assumed Japanese name, Yasumoto, until Son later persuaded the authorities to allow him to revert to his Korean surname. His father was a wheeler-dealer who lived by quick wits and the sweat of his brow, dabbling in bootlegged sake, swine farming, loan sharking, and later pachinko parlors—Japan’s ubiquitous low-stakes gambling game of clattering steel ball bearings.

“He was brought up pretending to be Japanese,” says Lionel Barber, former chief editor of the Financial Times and author of Gambling Man: The Secret Story of the World's Greatest Disruptor, Masayoshi Son. “So he’s living a lie. He’s undercover. He’s got two identities.”

At 16, Son read a business book by Den Fujita, who first brought McDonald’s to Japan. Fanatically inspired, he subsequently made 60 long-distance phone calls to beg Fujita for a meeting. After repeated rebuffs, Son bought a plane ticket to Tokyo and turned up uninvited at the McDonald’s head office. After haranguing the receptionist, he was eventually granted a 15-minute audience with Fujita, who advised his teenage devotee to focus on future technologies like computers. (Fujita later sat on the SoftBank board.)

Newly inspired, Son decided to move to the U.S., enrolling as a sophomore at a California high school and then studying economics at the University of California at Berkeley. Aside from his studies, Son dedicated five minutes every day to thinking about inventions, filling hundreds of notebooks. He eventually collaborated with some Berkeley tutors to invent the world’s first electronic translator, which he sold to Sharp Corporation. He then started a business importing second-hand arcade game machines from Japan.

His work ethic became the stuff of legend; after Son got engaged to a fellow Japanese student at Berkeley, he missed his own wedding twice due to being consumed by laboratory work (on the second occasion the judge took pity on his tearful bride and married them anyway, though they were way past their slot.) Despite now helming a successful American business, Son chose to return to his homeland, keeping a promise he’d made to his mother upon departing to the U.S.

In 1981, aged just 24, Son founded SoftBank, which originally served as a software wholesaler to support the burgeoning PC industry. (The name SoftBank initially meant a repository of software rather than anything to do with finance.) Just a year later TIME would name the computer its “Machine of the Year,” and Son was perfectly placed to cater to the new digital zeitgeist.

But soon after his return Son’s world was turned upside down when he was diagnosed with Hepatitis B and given three to five years to live. Now married with one young daughter and his second on the way, he underwent pioneering treatment that saved his life and endowed him with astonishing self-belief. “If you literally come back from the dead, that internalization of mortality is another huge motivating force,” says Barber. “He’s thinking: ‘I don’t have a lot of time.’”

A raft of tech-related ventures followed, including computer magazines and trade shows as well as memory chips. During the 1990s, Son invested $3 billion in 800 tech start-ups. In 1996, Son paid $100 million for 33% of Yahoo! Three years later, he sold off a chunk for a huge profit but still retained a 28% stake worth $8.4 billion. Son adopted a highly leveraged investment style—issuing SoftBank bonds to borrow money at rates cheaper than the banks.

It was the heady days of the dot-com bubble when Son’s net worth was surging by $10 billion a week. In February 2000, Son briefly unseated Bill Gates to become the world's richest person for three days. But when the bubble burst later that year, SoftBank shed 97% of its value, almost sending Son bankrupt. The $70 billion loss that Son personally suffered was at the time the single biggest implosion of net worth. Today, Son is sanguine about this “hiccup,” as he puts it.

“There was a bubble and crash, but that was not a crash forever,” he says. “It was the beginning of the Internet and most companies provided [services] for free. So there was not even a business model back then. But people's instinct was right. Now, everybody, every industry, uses the Internet.”

Indeed, also in 2000 Son took a $20 million bet on an obscure e-commerce startup called Alibaba—a stake that was worth $75 billion when the firm went public in 2014 and roughly doubled again by the time Son sold up. It stands among the most profitable investments of all time and helped more than anything to spin the legend about Son’s supposed Midas touch.

“Masa is a great thinker, his work ethic is phenomenal, nonstop, he’s a very, very driven guy,” Anthony Tan, founder of the SoftBank-backed ride-sharing platform Grab, told TIME in 2023. “He's got enough money, but he loves winning.”

Still, Son’s genius always had a reckless streak. Back in 2001, when he was struggling to get regulatory approval to set up a broadband subsidiary, he stormed into an official’s office at Japan's telecommunications ministry clutching a cheap cigarette lighter. “This is the end,” Son told the startled official, according to an interview he later gave with the Wall Street Journal, which confirmed the encounter with the official. “If you don't help me, I'm going to pour gasoline all over myself right here and set myself on fire with this $1 lighter.” Son got the help he needed.

After acquiring the Japanese subsidiary of Vodafone in 2006, the rebranded SoftBank Mobile emerged as a key player in Japanese telecoms. It’s a position largely owed to Son persuading Apple’s Steve Jobs to give him the exclusive rights to market the iPhone—history’s most successful consumer electronic product—when it debuted two years later.

Son’s foray into telecoms went global when he purchased Sprint in 2013. Son turned around the struggling U.S. provider and in 2020 pushed through a merger with T-Mobile that finally disrupted the AT&T and Verizon duopoly. Much as Son is typically a hands-off investor, leaving founders to do their thing, he proved himself operationally astute when required to roll up his sleeves.

“I vividly recall one meeting where he told the Sprint network team that he’d identified a blind spot in downtown Houston and how to fix it,” says Alok Sama, a former President and CFO of SoftBank Group International, and author of The Money Trap: Grand Fortunes and Lost Illusions Inside the Tech Bubble. “That was his level of detail.”

These were the headline successes that Son leveraged to create the SoftBank Vision Fund in 2017. With over $100 billion in capital, it’s the world's largest private equity fund. Son secured some $45 billion from Saudi Arabia’s Public Investment Fund (PIF) following a 45-minute meeting with Crown Prince Mohammed Bin Salman. “One billion dollars per minute,” Son bragged to David Rubenstein in a 2017 interview.

Son’s strategy was to invest a minimum of $100 million to juice each startup to market dominance by blowing competitors out of the water—a strategy that would be dubbed “blitzscaling.” In less than three years, the Vision Fund had burned through $76.3 billion in companies like Nvidia, Uber, WeWork, Paytm, Ola, and Flipkart. In 2019, SoftBank with great fanfare launched Vision Fund 2 with a touted value of $108 billion. However, it reportedly only managed to secure $30 billion, mostly self-funded.

This wasn’t helped by the dubious performance of the original Vision Fund, which in the 2021 financial year posted record losses of $27.4 billion as the value of tech stocks plummeted. External factors played a part: Russia’s war on Ukraine, the COVID-19 lockdown of financial megacity Shanghai, and Beijing’s crackdown on its tech giants—many of which were backed by SoftBank—combined to decimate investor confidence. 

Then there were the misses. In the late 1990s, SoftBank almost bought a 30% stake in Amazon for $100 million but backed out as Jeff Bezos believed his company was worth 16% more than Son’s valuation. That 30% would be worth $780 billion today. In 2017, Son held 4.9% of Nvidia—a sum that would be worth more than $200 billion today, though he cashed out in 2019 albeit for a $3.3 billion profit. “My heart is breaking!” Son laughs when asked about that decision. “I'm ashamed. I was very, very bullish.”

“I still want to buy a lot of Nvidia stock, but I also have to focus on my belief and support of OpenAI,” he adds. “We also have to make data centers, Stargate happen. If I had a bigger wallet, I would be investing more and more into Nvidia and a few other companies as well.”

Meanwhile, newer, more efficient AI models may make massive data centers obsolete, while strained energy grids could lead to high operational as well as environmental costs, undermining economic unviability. “We’re on a path where we may end up 40% overbuilt in terms of data centers,” says Paul Kedrosky, an investor and researcher at the intersection of technology and economics. 

Son disagrees. For over half-a-century, Moore’s Law has decreed that compute doubles every 18 months or so. But Son sees 10 times more AI chips being deployed in each three-year cycle, over which period the chips themselves are becoming 10 times more potent, while the AI models are also ramping up productivity by a factor of 10. “That’s 1,000x in three years,” says Son. “Nine years with three generations is 1,000,000,000x. It's a huge, huge difference.”

Still, critics caution that the collaboration between OpenAI, Oracle, and SoftBank could form a cartel that stifles innovation while inflating costs. Son is unapologetic about the technology coalescing around a few major players. “It will naturally happen,” he says. “For the AI race, it requires hundreds of billions of dollars of investment into the data centers, buying chips, integrating chips, training the models,” he says. “It's very, very costly, so it will naturally be concentrated into several very capable companies in terms of talent, capitalization.”

Son also shrugs off concerns that, even if successful, AI poses the risk of putting even more wealth and power into the hands of even fewer individuals. “In the Internet era, search is concentrated into Google; social networks are concentrated into Meta; E-commerce is concentrated into Amazon. But because they are very responsible companies—public companies, overseen by many regulators, shareholders, media, politicians—so they have to behave, and they do behave when they grow to some scale.”

The contention that the tech giants are responsible stakeholders has no shortage of rebuttals—even before their founders burrowed to the beating heart of the White House. Certainly, the breakneck pursuit of Stargate is driven in part by Washington’s desire to beat Chinese competitors in what’s framed as an AI “arms race.” Son—once a bridge between East and West—has now picked sides in this new era of Great Power competition. 

Still, Son is dismissive of the low-cost models offered by China’s AI pioneers, such as DeepSeek, when compared to the frontier research conducted by OpenAI. “Real innovation cannot happen just by copying and chasing,” he says. “You have to make real research, real developments, and that’s costly—but it has a big reward as first mover advantage.”

It’s an advantage Son hopes will be augmented by supportive regulation. On his first day in office Trump revoked a 2023 executive order by Joe Biden that sought to reduce the risks that AI poses to national security, the economy, public health, or safety by requiring developers to share the results of safety tests with the U.S. government before their public release. Son says regulation is “necessary … to maintain the healthy growth of the industry. We cannot let AI innovate in a way that is harmful to human beings.” However, “you have to have the right balance of supporting, accelerating innovation.”

Son is the consummate salesman, but it's important to separate the legend from the facts. Right from the beginning, his origin story and exploits have been carefully curated. While Son was indeed born amid the stench of pigs in a slum, his bootstraps ascension is largely a myth: by the time he left for college in the U.S., his family was so wealthy through pachinko parlors that he received a regular $2,500 stipend and drove around campus in a Porsche.

In California, Berkeley Prof. Forrest Mozer, Son’s first business partner, told Barber that he was “cheated” on the Sharp deal and never received a dime. (Son responded that there was only ever an informal agreement for an hourly wage, which Mozer concedes as true.) Son has also said he later made $1.5 million from importing arcade game machines, though other sources put the profit at $200,000.

What about Son’s greatest victory? He likes to take credit for unearthing Alibaba’s charismatic founder Jack Ma, highlighting he was the only CEO whom SoftBank invested in after meeting ten on a visit to China. “I can smell [Ma]. We are the same animal; we are both a little crazy,” Son told a forum in Tokyo in 2019. Perhaps, but it wasn’t a complete Hail Mary; Goldman Sachs had already backed Alibaba to the tune of $3.3 million.

When it comes to selling off his Nvidia stock, Son tells TIME the decision was forced “because of our financial constraints during COVID. We had to go into the protected mode.” In fact, Son sold up in January 2019—a full year before the pandemic struck.

“What you see isn’t always the reality,” says Barber. “Because he’s so good at playing the joker, but there's a calculating mind here. Even though he's got a big mouth and big vision, he loves to wait, he studies the enemy, studies the field, does his homework, loves to be underestimated. But when he strikes, he’s ruthless.”

But does he deliver? 

Stargate is already underwhelming. Of the total $500 billion to be spent over four years, some $100 billion was to be invested “immediately,” with the project touted to create 100,000 permanent jobs. However, only roughly $10 billion has so far been deployed in the Texas city of Abilene, where some 7,000 temporary construction jobs have been created, providing a bump to the local economy though also sparking a housing crisis. Investor interest has been tepid. In January, SoftBank backed away from talks to acquire U.S. data center operator Switch for $50 billion.

With something like 70% of all venture capital going to AI companies, there are consequences if the AI bet fails. “It’s really important to look at the other side—who's not getting money,” says Kedrosky. “Small manufacturers have not been getting money. Other technology companies doing interesting things in life sciences [find it] very hard to raise money right now. So there's a host of people who are not getting money because it’s all going to this one place.”

Besides, while the telecoms bubble of 2000 left in its wake fiber optic cable and 3G infrastructure that proved invaluable over the following decade, data centers—essentially warehouses piled high with GPUs—don’t have the same longevity. Even the smartest tech minds are cautioning that we need to see real applications generating revenue rather than capacity being senselessly built up.

“For this not to be a bubble, by definition it requires that the benefits of this are much more evenly spread,” Microsoft CEO Satya Nadella said at the World Economic Forum in Davos last month. “I think a telltale sign of if it’s a bubble would be if all we’re talking about are the tech firms.”

The fear is that Son’s bullishness on AI is based on wishful thinking. When Son first approached Saudi Arabia about investing in the Vision Fund, the Kingdom’s Commerce Minister asked him: “You went from zero to hero to zero. How do we know you won’t go back to zero again?” Sure enough, Son did indeed suffer another precipitous fall, though today is once again brimming with braggadocio. Though it’s his gambler’s propensity for lurching between boom and bust that looms over Stargate and American AI ambitions in general.