IF YOU want to find a spectacular vision of the future, Silicon Valley is not the only place to look. In Tokyo Masayoshi Son, the boss of SoftBank, a Japanese telecoms group, is starting an investment fund worth $100bn which, he hopes, will make him the Warren Buffett of technology. “Masa” is no stranger to risky bets: SoftBank was an early investor in Alibaba, a Chinese e-commerce company, and has sunk $22bn in Sprint, a struggling American telecoms firm. Now he has been seized by the kind of Utopian fever that would make the Sage of Omaha choke on his Cherry Coke.

Mr Son, who is 59, believes that the world will soon encounter what is known as the Singularity, the point at which artificial intelligence exceeds the human kind. The brains of people and machines will become enmeshed (see article). Every person will have over 1,000 devices linked by a seamless global network, with the data analysed by machines in the cloud. As well as smart glasses, people will wear smart shoes and every car and washing machine will link up to the web. This internet revolution, says Mr Son, will be more momentous than the first.

He has begun making acquisitions. Last year he spent $31bn buying Britain’s ARM Holdings, which designs the chips in mobile devices (it will be owned jointly by SoftBank and the fund). He also invested a total of $2bn in OneWeb and Intelsat, two satellite-technology firms that aim to launch thousands of microsatellites to orbit the Earth providing high-speed internet access. Tech firms around the world are bracing for more swoops by Mr Son, who says his aim is to build a business empire lasting 300 years. What he doesn’t mention is that he also wants to prove beyond all doubt that his fortune is due to skill, not one lucky deal.

Mr Son believes he has anticipated successive paradigm shifts in technology. The son of an ethnic-Korean pig farmer, whose childhood was spent in a shack in southern Japan, Masa wept joyfully when, as a teenager, he first saw a picture of a microchip. He learned programming while at the University of California, Berkeley, then in the 1980s sold software in Japan. He was an early investor in internet firms, buying a share of Yahoo in 1995 and the Alibaba stake in 1999. Later he invested in mobile telecoms, first in 2006 with his purchase of Vodafone’s Japanese mobile arm and then of Sprint in 2013. Now SoftBank is huge, with an enterprise value (its market value plus its net debt) of $193bn.

Yet Mr Son’s career is still defined by Alibaba. In 1999 he was visited in Tokyo by Jack Ma and Joseph Tsai, co-founders of a fledgling website in Hangzhou. Mr Son tapped on a calculator as they haggled and agreed that SoftBank would buy 30% of the young firm for $20m. The deal was “based on my sense of smell”, Mr Son said later. Now Alibaba’s market value is $270bn, and, after selling some shares last year, SoftBank still owns 28%.

About 95% of SoftBank’s market value is accounted for by the Alibaba stake, so the rest of what it does, from telecoms to venture capital, may be worth little, once debts are deducted. Mr Son says that SoftBank has made an internal rate of return of 43% on all its other investments, excluding Alibaba, but the basis of his calculations is unclear. There have been triumphs—SoftBank made $5bn buying and selling Supercell, a Finnish gaming firm, between 2013 and 2016. But the group has produced little cashflow, and Mr Son’s deals have left it with $110bn of net debt.

So Mr Son has a minority investment in a great firm, but has yet to build one himself from scratch. And SoftBank’s poor finances are impeding his ambitions. Because his stake is only 19%, he cannot raise cash by selling shares without weakening his grip on the firm. He could sell the rest of the Alibaba stake, but appears reluctant to let go altogether. Or he could try to broker a merger of Sprint with T-Mobile, another American telecoms firm, allowing SoftBank to rid its balance-sheet of Sprint’s $31bn of net debt. Until now antitrust regulators have opposed a deal. But Mr Son hopes that the Trump administration will be more amenable.

The alternative is partially to bypass SoftBank, which is what the new $100bn fund achieves. Mr Son will have more discretion over what to buy, free of grumbling public shareholders. Outside investors will give him huge firepower. Saudi Arabia’s public investment fund, for example, has promised to give him buckets of cash. The fund and its debts will be kept off SoftBank’s books.


Investors in the new vehicle and owners of SoftBank shares should have three worries. First, while Mr Son’s ideas stand out for their intensity, they are not entirely original. Others in tech share his vision of ubiquitous, web-linked devices with their data crunched by machines, so the values of firms involved in these areas are sky-high; SoftBank paid 71 times earnings for ARM. Second, Mr Son can lose focus. Some of the startups he particularly admires, such as Uber and Airbnb, are only loosely related to his notion of the internet. Others are even more tangential. On March 20th SoftBank bought a $300m stake in WeWork, a trendy office-rental firm with a dizzying valuation.

The third worry is governance. Mr Son’s mind skips from one obsession to the next. In 2014-15 he was briefly infatuated by India’s tech scene, for example, and appointed Nikesh Arora, an Indian-born former Google executive, as his heir apparent, only to ease him out a year later, in 2016. It is clear that one man with a messianic streak will dominate the fund as well as the running of SoftBank. Mr Son’s dual role also produces conflicts of interest: if there is a juicy deal, who benefits—the fund or the firm?

For Mr Son, these are quibbles that will fade into irrelevance over his 300-year horizon. He has said that, looking back on his first six decades, he regrets that he “focused too much on the daily routine and didn’t really think big.” So far only 3% of his brainpower has been devoted to big investment decisions, he believes. Now more than half of his mental capacity will be directed at fulfilling his destiny. Masa is just getting started.