EVERY DAY, from 8am to 10pm, Son Masayoshi sits in his mansion in Tokyo doing what entertains him like nothing else: sizing up technology entrepreneurs and handing out money. Working from home has not slowed down the billionaire boss of SoftBank. At the Japanese group’s earnings call on May 12th Masa, as he is universally known, boasted of backing 60 companies in three months. Between January and March he doled out $210m a week.

In the past four years SoftBank has poured $84bn or so into startups. It was the world’s biggest tech investor even before complementing a $98.6bn investment vehicle it runs, the Vision Fund, with a sibling that now contains $30bn. The 224 tech firms it has backed range from early-stage startups to established giants like ByteDance, owner of TikTok, a Gen-Z time-sink. Names like Plenty, Better and Forward imply missions to transform industries such as food, health and banking. Going by the valuations used by SoftBank and other venture-capital (VC) funds, the companies backed are collectively worth a colossal $1.1trn, according to PitchBook, a data provider.

In spring 2020 SoftBank’s entire tech edifice nearly came tumbling down. As covid-19 spread and markets convulsed, SoftBank’s lenders took fright. Yields on its bonds surged. Investors wondered how—or if—Mr Son’s punts might weather the pandemic. His signature approach—big bets based less on spreadsheets than on “feeling the force” of a deal—looked riskier than ever. The initial public offering (IPO) of WeWork, an office-sharing firm, collapsed in late 2019, before the pandemic. Between February and March 2020, SoftBank’s share price plunged by over 50%.

“Masa got close to the flame,” sums up a person close to SoftBank. But then the markets rebounded. The Federal Reserve was pumping in liquidity by supporting the market for junk-rated corporate bonds (SoftBank’s debt is rated below investment-grade). SoftBank announced a sale of $41bn of its $252bn total assets to shore up its position.

Now Mr Son is once again looking like a genius. Backing tech darlings proved the perfect strategy in the digitally accelerated coronavirus economy. His hunches have paid off. In just over a year SoftBank has gone from survival mode to spewing out cash like a giant ATM. The listing of Coupang, a South Korean e-merchant, reaped SoftBank $24bn in profits. Several more firms it has backed launched into an IPO market that turned red-hot last summer. Last month SoftBank reported an annual net profit of $46bn, the highest ever by a Japanese company. And more is to come: on June 10th Didi Chuxing, a Chinese ride-hailing firm one-fifth owned by SoftBank, said it would list its shares at a valuation of around $100bn. Despite a recent dip, SoftBank is worth a cool $126bn, $60bn more than at the trough in March 2020.

As one big SoftBank shareholder puts it, after a spell like this, “they are untouchable.” Forgotten was the fact SoftBank’s bumper profits followed one of the largest losses in Japanese corporate history a year earlier. Forgotten, too, was a familiar criticism of recent years: that Masa, who is ultimately responsible for every yen spent by the group, picks hyped investments.

Perhaps most forgotten of all is that SoftBank, having come to epitomise the tech wave, has occasionally acted in what some onlookers regard as questionable ways. It has eschewed common governance standards. It has become entangled in Europe’s two biggest corporate scandals in recent years: over Wirecard, a fraudulent German payments processor, and Greensill, a British supply-chain-finance firm now facing bankruptcy. And it has fused ever more closely with its founder.

Created by Mr Son in 1981 to distribute computer programs, SoftBank first delved into internet services in the 1990s before reinventing itself as telecoms business. It bought the Japanese activities of Vodafone, a British mobile carrier, in 2006. In 2013 it bought Sprint, an American mobile provider. Along the way, in 2000, Mr Son paid $20m for a chunk of Chinese e-commerce upstart called Alibaba. That genius bet—Alibaba is now a global giant worth nearly $600bn—earned Mr Son kudos as a tech visionary. It also inspired him to transform SoftBank into an investment firm.

In 2018 Mr Son spun off some Japanese telecoms assets and unveiled the Vision Fund. Not content with raising a typical $1bn-10bn VC vehicle, with the help of Rajeev Misra, a well-connected ex-Deutsche Bank financier and Mr Son’s key lieutenant, SoftBank raised $45bn from Saudi Arabia’s Public Investment Fund (PIF) and $15bn from Mubadala, Abu Dhabi’s sovereign-wealth fund. The Japanese firm put in $28bn of its own cash and assets, including a slice of Arm, a British microchip designer it had bought in 2016.

Today SoftBank is best understood as having four main parts. The most valuable by far is the 24.85% stake in Alibaba, worth $144bn. The second part contains its remaining mobile businesses and Arm (with Arm heading for disposal). SoftBank’s two Vision Funds make up the third portion. Finally comes Northstar, an internal hedge fund set up a year ago. At the centre sits Mr Son. “It’s one company, the Masa show, that’s it,” sums up a former executive.

Mr Son has two big ideas for the latest iteration of SoftBank. The first is to combine tech investing with financial engineering. In order to juice up returns, he has taken traditional VC and piled on leverage and complex structures. Second, Mr Son wants to create an over-arching tech “ecosystem” with SoftBank at its heart. Although SoftBank often owns only slivers of companies, he wants them to work together as if part of one group. The model is evocative of, if not identical to, the keiretsu—Japanese conglomerates such as Mitsubishi, with tentacles in finance, carmaking and lots besides, all helping each other.

It’s complicated
Start with the financial gymnastics. At SoftBank’s highest levels, close to Mr Son, is a group of traders who worked at Deutsche Bank at a time when the German lender was famous for its appetite for risk-taking. Chief among them is Mr Misra, who heads SBIA, the part of SoftBank running the Vision Fund. “There are people at the Vision Fund whose entire raison d’être is financial engineering,” says a person who knows SoftBank well. “Complexity is their best friend—if they want to get from A to B they will go through all the letters of the alphabet to get there.” Mr Son appears to prefer their swashbuckling ways to those of old-school, more conservative VC types.

It was such ways that brought SoftBank into the Wirecard fiasco. In early 2019 the German firm’s share price had fallen by half from its peak in late 2018 after reports of accounting irregularities. On March 28th 2019 Wirecard, by then a member of Germany’s blue-chip DAX 30 index, said it was suing the Financial Times over a series of investigative reports. At that moment another ex-Deutsche Bank trader at SoftBank, Akshay Naheta, chose to back the controversial company with the full weight of his employer’s reputation.

On April 24th 2019 Wirecard announced that a SoftBank “affiliate” would invest €900m ($1bn) via a convertible bond. SoftBank also struck a co-operation agreement with Wirecard that opened the way for the payments firm to do business with other SoftBank companies, including those in the Vision Fund. The apparent validation of Wirecard by the world’s largest tech investor sparked a 21% rise in the German firm’s share price. It also bolstered Wirecard’s creditworthiness.

The Wirecard deal had two unusual aspects. For a start, it turned out to be ephemeral. It later emerged that the €900m bond was subject to a refinancing exercise almost as soon as it was issued, reaping an immediate €64m profit. When Wirecard went bust in June 2020, after the FT’s reporting proved accurate, it was other investors who lost out.

Another oddity was that SoftBank itself put no money into Wirecard. The “affiliate” fund that did was managed by SBIA, the Vision Fund overseer. Its investors included a small group of individual SoftBank executives aiming to make a personal profit, among them Mr Naheta, Mr Misra and Sago Katsunori, SoftBank’s former strategy chief, and one of Abu Dhabi’s sovereign-wealth funds. Wirecard publicised its agreement with SoftBank to financial markets, an association that for SoftBank carried reputational risk given the FT’s reporting on Wirecard. In the end the bulk of the immediate profit on the bond trade went to the affiliate’s individual investors, not to SoftBank or its shareholders. Wirecard’s later implosion confirmed the reputational risk that SoftBank had run. A person familiar with the matter also confirmed that the position of the SBIA-managed affiliate fund meant it stood to lose potential gains if Wirecard went bankrupt.

One person in the know describes Mr Naheta and colleagues as putting a team to work to introduce Wirecard to certain Vision Fund portfolio companies, in accordance with a co-operation agreement between SBIA and Wirecard. These portfolio firms included stars of the tech world, whose tacit endorsements could have given the ailing German company a public-relations fillip—possibly boosting the SoftBank executives’ returns. In the event, Vision Fund firms avoided Wirecard owing to the FT’s reporting.

Northstar is another example of Mr Son’s embrace of high finance. SoftBank’s newest unit looks like the exact opposite of Mr Son’s self-proclaimed desire to back exciting new firms and think in terms of an investment horizon stretching 300 years into the future. Northstar’s mandate seems to be to make short-term bets on public stocks that anyone with access to a brokerage account could buy or sell. One reason for this apparent departure from Mr Son’s guiding philosophy was that SoftBank had cash lying around. When the financial pinch of the early pandemic eased, some proceeds from the $41bn of asset sales could be reinvested. Most companies park spare cash in dull securities like government bonds. But that is not the Masa way.

Instrumental to Northstar is Mr Naheta, who runs the unit. The 39-year-old was not the obvious pick to head a fund placing some of the world’s biggest stockmarket bets. After leaving Deutsche Bank he set up his own stock-picking firm, which seems to have focused on mid-cap stocks. Northstar, by contrast, has deployed huge sums of money, using SoftBank’s spare cash magnified by leverage. Last September investors started noticing that one market player’s tech-focused bets were so big (and structured in such a complex way) that it single-handedly caused some companies’ share prices to rocket. It was not long before Northstar was unmasked as the “Nasdaq Whale”. It lost $5.6bn on derivative transactions in the year to March 2021.

In what appeared to be a letter in response to a freedom-of-information request by PlainSite, an investigative news outlet, the Securities and Exchange Commission said in March it was investigating SoftBank. A SoftBank spokesperson says the company is “not aware of any SEC investigation into the company’s securities trading and has not been notified of such”.

Techno-ecology
If Wirecard and Northstar exemplify Mr Son’s penchant for financial gymnastics, the story of Greensill’s rise and fall highlights the dangers of its ecosystem. SoftBank touts this as a competitive advantage. It regularly hosts get-togethers for all the founders whose firms it has backed. At its best the ecosystem can be a way for entrepreneurs to share ideas, bolster sales and boost prospects. But it can also lead companies in which SoftBank is a minority shareholder to feel the need to favour other parts of the ecosystem—even if this is not obviously in the interest of non-SoftBank investors or counterparties. SoftBank says its portfolio firms “have full autonomy to decide whether or not to work together”.

For SoftBank Greensill could serve multiple purposes. Its Australian founder, Lex Greensill, had a well-rehearsed story about using tech to transform a stultified industry—Masa’s special formula. Greensill could be used for the benefit of SoftBank’s ecosystem. Its flagship product involved making loans to tide companies over after they issued invoices to customers but before they received payments. Greensill then repackaged these invoice-backed loans and sold them to investors. Credit Suisse, a bank, offered to pitch funds full of those bonds to clients, including family offices and corporate treasurers.

Mr Son called Mr Greensill the “money guy”. Starting in May 2019 the Vision Fund invested $1.4bn in Greensill, turning its founder into a paper billionaire. Early last year the investment started proving its value. Some companies in the Vision Fund badly needed money. Firms like Katerra, a now-bankrupt American construction startup, found that building houses could not easily be made cheaper and faster with software. Oyo, an Indian hotel group, tried to expand too fast. After the WeWork debacle potential lenders steered clear of many Vision Fund firms. Greensill could help fill the hole. It lent money to Katerra and Oyo—or, rather, Credit Suisse clients did so indirectly. At times it was not in SoftBank’s interest for some of its portfolio firms to raise fresh equity: a lower valuation might oblige SoftBank to revalue its shareholding, generating a loss. Having Greensill on hand to extend loans to struggling firms proved useful in the end.

This introduced potential conflicts of interest. SoftBank invested in the company that made the loan (Greensill) and in those it lent to (Katerra, Oyo and others). SoftBank invested over $500m in Credit Suisse funds. Greensill’s demise means its erstwhile backer can paint itself as one of its financial victims. A former SoftBank executive familiar with the matter says the situation was complex but that conflicts of interest could be managed. Credit Suisse, for one, is unimpressed. It is preparing for litigation against SoftBank. SoftBank says, “Any potential conflicts were appropriately managed in accordance with existing SBIA policies.”

In late 2019 it seemed that SoftBank might take a different, less controversial path. SoftBank’s shares were trading at an estimated 70-75% discount to the value of the company stakes it owned, a historic low on that measure. Elliott Management, an activist hedge fund from New York, spotted an opportunity. Simplifying SoftBank’s structure, improving governance and returning money to shareholders were all time-tested ways to cause the conglomerate discount to narrow. The early-pandemic scare forced Mr Son’s hand. The $41bn of asset sales included most of SoftBank’s remaining American telecoms businesses. It sold Arm to Nvidia, a bigger chipmaker (the deal awaits regulatory approval). This streamlined the group somewhat. But it also accelerated the metamorphosis from a juiced-up telecoms utility into a complex investment holding company.

Elliott’s prodding prompted some standard-issue governance improvements, such as appointing a woman to its all-male board and lifting the number of independent directors. As of June 2021, four out of nine board directors count as independents, up from three out of 12 in January 2019. Yet few observers think this resolves all potential problems. Since the start of last year SoftBank has suffered departures of highly experienced senior executives involved in legal affairs and compliance. A shared concern, according to a person close to some of them, was SoftBank’s culture. The firm is alleged to be permissive of conflicts of interest. The use of financial intermediaries was another concern.

The latest person to announce her departure was Kawamoto Yuko, SoftBank’s first female director and a respected corporate-governance expert. She left after just a year to take a job as a commissioner at Japan’s National Personnel Authority, having reportedly disagreed with Mr Son over internal controls. On May 21st she published a description of governance at SoftBank, calling for more internal checks and more dissenting voices. She commented that it would be nice if the “obligation to dissent” or to disagree when necessary, was more widespread throughout SoftBank. SoftBank says that “constructive debate is the sign of an effective board” and that Ms Kawamoto agreed with governance changes including the appointment of a chief risk officer, which she had suggested. “She did not leave because of a disagreement, but rather because she was appointed to a government position,” a SoftBank spokesperson says.

Under the existing governance arrangements, potentially problematic dealings can crop up. In return for facilitating the SoftBank affiliate’s convertible-bond deal, for instance, Wirecard and SBIA paid multimillion-dollar success fees to a German financier, Christian Angermayer, according to the FT. In another example, last summer Marcelo Claure, SoftBank’s chief operating officer, bought 5m shares of T-Mobile, a mobile operator, for around $500m. The purchase was funded by a loan from SoftBank. T-mobile’s share price has risen, and Mr Claure will keep the upside. But had it fallen, SoftBank would have had to find a way of get hundreds of millions of dollars back from its own executive. A SoftBank spokesperson says “the loan to acquire T-Mobile shares further aligns the interests of SoftBank shareholders with management. Under the terms of our merger agreement, SoftBank stockholders will receive over $7bn in T-Mobile equity if the company continues to perform and achieves a stock price of $150.”

According to a person familiar with the Vision Fund, in at least one case an executive invested privately in an unlisted firm before SoftBank backed it, resulting in a large valuation increase. Deep Nishar, senior managing partner at SBIA, made a personal investment in Petuum, an artificial-intelligence startup founded in 2016, before leading the Vision Fund’s $93m bet on the firm in 2017. SoftBank’s policies let Mr Nishar keep the personal investment, which was disclosed to the Vision Fund’s limited partners. SoftBank says that “the investments in Petuum were disclosed, complied with the firm’s policies, and are relatively common practice in growth-stage investing.” Other VC firms call such a practice rare, and typically require executives to sell equity stakes in a company at cost to their firm in such situations.

Another big question governance experts have wrestled with is where SoftBank ends and Mr Son’s individual interests begin. The boundaries are seemingly not always clear. Some of the firm’s activities are designed so an outsized share of profits flows to Mr Son relative to other SoftBank shareholders. Take Northstar. When the fund was set up, a third of the money invested was Mr Son’s; that brings him a third of the profits generated by Northstar. But the fund—and so Mr Son personally—benefit from belonging to SoftBank, 71% of which is owned by other shareholders. Its trades are either explicitly or implicitly backed by SoftBank’s balance-sheet.

An obvious potential problem with a CEO having a personal interest in a particular division of his firm is that such a boss cannot neutrally allocate capital. Diverting cash to Northstar, in this case, can lead to its profits surging—and flowing in part to Mr Son. Northstar’s structure, including Mr Son’s personal stake in it, was approved by SoftBank’s board, which discussed the matter independently of Mr Son.

Not so loanly
SoftBank insists its governance and finances are sound. It has long had a policy of keeping enough money on hand to repay all its bonds that mature in the next two years. The asset disposals after last year’s crisis resulted in a less indebted group. SoftBank says it does not want to borrow more than a quarter of the value of the holdings (for example in Alibaba) that are often used as collateral for the loans.

This is only a partial comfort to minority shareholders and other observers. SoftBank’s professed leverage cap is high by investment firms’ standards. Standard & Poor’s (S&P), a credit-rating agency, says it disagrees with how SoftBank calculates the ratio of credit to assets. SoftBank says that S&P calculates the ratio using its own methodology; it has revised its rating outlook for SoftBank from negative to stable. Some of the investment holding companies rated by S&P receive higher ratings than SoftBank even though their ratios after S&P’s adjustments are weaker than SoftBank’s. A SoftBank spokesman says that as it establishes a record of improvement this will help S&P’s ratio level. The company is “continuing our communication with [S&P] for upgrading”.

Borrowing pops up across the company. The Vision Fund can borrow against companies it owns, which are themselves indebted. Mr Son is known to have pledged his own shares in SoftBank to fund activities related to the firm. Moody’s, another credit-rating firm, describes the resulting capital structure as fluid, complex and having limited transparency. Analysts gripe disclosure is patchy at best.

Corporate governance, debt, SoftBank’s association with Wirecard and Greensill: neither is likely to be troubling Mr Son right now. The signs are he feels emboldened after surviving the corporate version of a 100-year flood. This carries risks, warns a SoftBank shareholder: “Masa and the senior team have been so successful, there is a feeling of ‘can everyone just leave us alone’…That could be dangerous if they think there is less need to take notice of corporate governance.” Investors were surprised by the news in May that instead of continuing to buy back shares, as they would like, Mr Son is tripling the size of Vision Fund 2, from $10bn to $30bn. Unlike the first Vision Fund, where outside limited partners sometimes acted as a brake, the new fund has no external investors.

Masa seems to have been given even freer rein to feel the force. On June 10th the new fund led a $639m investment in Klarna, a Swedish fintech firm. That and other similar-sized bets are much larger than last year’s investments of under $100m. SoftBank has set up three special-purpose acquisition companies hoping to raise a combined $1bn or so, and then merge with startups—possibly including, SoftBank has said, some in the two Vision Funds. That would kick up still more potential conflicts of interest.

The bull case for SoftBank is simple. It is an unabashed wager on tech-fuelled firms continuing their meteoric rise. It can thrive as long as investors are on hand to fund loss-making companies in the hope of future riches. For now, they are. But the recent IPO boom is petering out. Much of SoftBank’s record profit came with an asterisk: the share prices that helped generate it had already started falling back to Earth. Coupang has lost a fifth of its market value since listing. SoftBank and its shareholders are aware the party in the markets could come to an end, especially if central banks raise the ultra-low interest rates that make borrowing cheap and growth stocks appealing. Masa’s rebound last year was swift—but also lucky. The next stress test for SoftBank may not be far off.