Gunjo · Business Intelligence for the AI Era
← Sticker Wall JOURNEY · DETAIL

SoftBank's Masayoshi Son: From Software Distribution to the $100 Billion Vision Fund Journey

Founded: Masayoshi Son · SoftBank Group Corp.

JOURNEY

Key Fields

FIELD STAMPS
IndustryConglomerate / Trading House
RegionJapan
ScaleGiant
ChannelHybrid

Origin

Masayoshi Son was born in 1957 in Saga, Japan, to a Korean-Japanese family. As a teenager, he was inspired by Sakamoto Ryoma and harbored ambitions of entrepreneurship. In the 1970s, he moved to the U.S. to study economics and computer science at UC Berkeley, earning his first $1 million through a patent for a pocket translator. In the early 1980s, as the Japanese PC industry was nascent and software distribution was highly fragmented, he identified software as the core infrastructure of the information age and decided to return to Japan to start a software wholesale business. On September 3, 1981, at the age of 24, Son founded SOFTBANK Corp. in a small office in Tokyo. On his first day, he told his only two employees that he aimed to build a company with 1 trillion yen in revenue. Both employees resigned the next day, but driven by his conviction in the future of the software industry, Son eventually built that small shop into Japan's largest technology conglomerate.

Milestones

1981
Software Distribution Growth
Founded SOFTBANK Corp. in 1981, focusing on wholesale distribution of PC software, quickly becoming Japan's largest software distributor. Entered publishing in 1982, launching magazines like Oh!PC and Oh!MZ, eventually becoming Japan's largest computer technology publisher with individual magazine circulations reaching 140,000 by 1989. Listed on the Japanese over-the-counter market in 1994 with a $3 billion valuation. At 37, Son experienced the leverage of capital markets for the first time, setting the stage for his later aggressive investment strategy. This phase lasted from 1981 to 1994.
1995
Internet Bets Turning Point
In 1995, spent $800 million to acquire the COMDEX computer trade show and $2.1 billion to acquire Ziff Davis publishing, going all-in on the internet bubble. In 1996, partnered with Yahoo to launch Yahoo Japan, which became the country's top portal. His biggest failure was the $1.5 billion purchase of an 80% stake in memory manufacturer Kingston Technology in 1996; by 1999, due to a market crash, he sold it back to the original owners for one-third of the price, losing about $1 billion. This was Son's largest single loss at the time, leading him to adopt the philosophy that 'betting on the right track is more important than betting on a single company.' This phase lasted from 1995 to 2001.
2000
Investing in Alibaba Pivot
In 2000, near the peak of the internet bubble, Son invested $20 million in a nascent Chinese e-commerce company, Alibaba, for a ~30% stake, following a 6-minute meeting with Jack Ma in Beijing. Although SoftBank's own valuation plummeted 99% from $180 billion to less than $2 billion after the bubble burst, and Son's net worth briefly hit zero, the Alibaba seed grew into a $60 billion return when it went public in 2014. It became one of the most successful single VC investments in history and provided the foundation for SoftBank's later confidence in launching massive funds.
2006
Telecom Expansion Growth
In March 2006, acquired Vodafone Japan for approximately $15.1 billion, rebranding it as SoftBank Mobile. In 2008, secured exclusive rights to the iPhone in Japan, transforming from a software investment firm into Japan's third-largest wireless carrier. In 2013, acquired Sprint, the third-largest U.S. carrier, for $22 billion, marking SoftBank's expansion into a multinational telecom group. However, the Sprint deal suffered long-term losses, and after its merger with T-Mobile in 2020, SoftBank's stake was diluted. The dream of a telecom empire did not go as planned. This phase lasted from 2006 to 2013.
2016
Acquiring Arm Pivot
In 2016, following the Brexit referendum and the subsequent crash of the British pound, Son seized the opportunity to acquire UK chip designer Arm Holdings for over $32 billion. It was the largest single transaction in SoftBank's history and the largest tech M&A in Europe at the time. Son's bet was that IoT and mobile chips would dominate the next decade, with Arm architecture permeating everything from phones to cars and servers. After the acquisition, Arm's performance fell short of expectations, and combined with SoftBank's debt pressure, the company considered selling or listing Arm between 2018 and 2020, ultimately choosing to hold it long-term.
2017
The $100 Billion Fund Turning Point
Partnered with Saudi Arabia's sovereign wealth fund, PIF, to launch the SoftBank Vision Fund, raising $93 billion. SoftBank contributed $28 billion, while Saudi PIF committed $45 billion over 5 years as the largest LP. Apple, Qualcomm, Foxconn, Sharp, and Mubadala also joined, creating the largest tech venture fund in human history. That same year, the $108 billion Vision Fund 2 was proposed. Son declared he would use the fund to buy '300 companies of the AI revolution,' but no investor could truly manage a $93 billion VC fund, and the outcome proved this ambition far exceeded reality.
2017
Aggressive Investing Growth
In March 2017, invested $300 million in WeWork, followed by an additional $4.4 billion in August to solidify its position in the co-working sector. In 2017, spent $10 billion to bet on Uber, becoming its largest shareholder. In March 2018, led a $535 million investment in DoorDash. During the same period, invested in a massive list of unicorns including OYO, Opendoor, Grab, Paytm, Flipkart, ParkJockey, Coupang, and Auto1, spanning ride-hailing, food delivery, hotels, e-commerce, parking, and real estate. The pace, with investments ranging from hundreds of millions to billions of dollars, shocked Silicon Valley and Wall Street, earning the nickname 'war with a checkbook.' This phase lasted from 2017 to 2019.
2019
WeWork Collapse Failure
In 2019, after WeWork filed its IPO prospectus, corporate governance scandals erupted. Founder Adam Neumann was ousted, the valuation plummeted from $47 billion to $2.9 billion, and the IPO was withdrawn. SoftBank had invested over $10 billion, was forced to inject an additional $1.5 billion, and initiated a $3 billion stock buyback to save the company. In November 2023, WeWork filed for bankruptcy, and SoftBank's investment was effectively wiped out. This was the most glaring failure of Son's investment career; he admitted at an earnings call that his judgment on WeWork was wrong and his blind trust in Adam was a mistake.
2020
Pandemic Collapse and Self-Rescue Failure
Due to the COVID-19 pandemic and the oil price crash, SoftBank's stock price fell to the 3,000 yen range, and its market cap halved. Son urgently launched a 4.5 trillion yen (approx. $41 billion) asset sale plan to raise cash for a 45% stock buyback. In September of the same year, it was revealed that SoftBank had bought massive call options on tech stocks, betting on giants like Apple and Amazon, earning the nickname 'Nasdaq Whale' from Wall Street. After losing billions, they liquidated their positions, further damaging their reputation. In 2020, the Vision Fund reported a 1.8 trillion yen annual loss, its first since inception three years prior.
2020
Aborted Arm Sale to Nvidia Failure
In 2020, announced the sale of Arm to Nvidia for $40 billion, which would have been the largest semiconductor M&A in history. However, the deal faced antitrust scrutiny from the U.S., U.K., EU, and China, and the Arm China joint venture refused to cooperate due to internal equity disputes. In February 2022, both parties officially abandoned the deal. The cost of this failed transaction was two years of strategic stagnation and massive legal and investment banking fees, but it unexpectedly allowed SoftBank to retain full ownership of Arm, setting the stage for the 2023 turnaround.
2023
Vision Fund Massive Loss Failure
Financial reports revealed that the Vision Fund's annual loss widened by 70% to $32 billion, the largest record since its inception. During the same period, SoftBank was forced to sell Fortress Investment Group to Mubadala and sell its early-stage investment arm, SoftBank Ventures Asia, to Masayoshi Son's brother's company to cut costs. Valuations of many portfolio companies, including OYO, Katerra, Zume, Getaround, and Rappi, shrank by over 70%, leading to layoffs and closures. The outside world began to believe the SoftBank VC experiment had completely collapsed, and Son himself began to tone down his high-profile 'AI revolution' rhetoric in public.
2023
Arm IPO Turnaround PMF
In 2023, Arm went public on the Nasdaq, raising $4.87 billion at a $54.5 billion valuation, with SoftBank retaining a 90.6% stake. Before the IPO, SoftBank repurchased a 25% stake in Arm from the Vision Fund for approximately $16 billion to pave the way. Arm's stock surged 25% on its first day of trading, and as the AI boom drove the penetration of Arm architecture in data centers and edge computing, the stock price continued to climb, nearly doubling by early 2024. This became the key turning point for SoftBank to emerge from its darkest hour (2019-2022), restoring market confidence in Son's long-term judgment.
2025
All in AI Infrastructure Turning Point
In December 2024, announced a $100 billion investment in the U.S. over 4 years to create 100,000 jobs. In January 2025, partnered with OpenAI, Oracle, and MGX to launch the $500 billion Stargate AI infrastructure project. In February 2025, established the SB OpenAI Japan joint venture, spending $3 billion annually to deploy OpenAI solutions. In March 2025, acquired Ampere Computing for $6.5 billion to enter the ARM server chip market. In October 2025, acquired ABB's robotics business for $5.375 billion. By the end of 2025, cumulative investment in OpenAI reached $34.6 billion, holding approximately 11% to become the third-largest shareholder. Arm's stock price hit a 2024 high, and SoftBank transformed from the 'most miserable VC' into the biggest infrastructure gambler of the AI era. This phase lasted from 2025 to 2026.
2026
Profitability Recovery Growth
In February 2026, SoftBank's quarterly report showed a return to profitability. During the period, it had invested over $30 billion in OpenAI for an 11% stake and was in talks to invest another $30 billion in a funding round valuing the company at $750 billion to $830 billion. In May 2026, announced an investment of up to 75 billion euros in France to build a 5GW AI data center, with the first phase of 45 billion euros located in the Île-de-France region. PayPay also went public in the U.S. in March 2026. From the ruins of the Vision Fund to an AI infrastructure empire, SoftBank completed a total narrative reconstruction in 5 years, though the scale of the bet is an order of magnitude larger than the original Vision Fund.

Turning Points

  • In 2000, during the darkest hour when the internet bubble burst and SoftBank's stock plummeted 99%, Son insisted on not selling Alibaba. The $60 billion return 14 years later established SoftBank's DNA for making massive, long-term bets.
  • The 2016 acquisition of Arm for $32 billion during the post-Brexit pound crash became the ticket for shifting from a VC gambler to a chip infrastructure player, and serves as the underlying asset for the 2023 turnaround and 2026 AI narrative.
  • The 2019 WeWork IPO collapse and the 2020 'Nasdaq Whale' scandal forced Son to admit his investment methodology was flawed, leading him to shift from 'checkbook shopping' to focusing on a few core tracks.
  • The 2023 Arm IPO allowed SoftBank to climb out of the loss mire, proving the value of Son's long-term holding strategy and providing the capital and confidence for his 2025 'All in AI' move.

Failures & Pitfalls

  • The WeWork investment, which saw over $10 billion evaporate, is SoftBank's most public and glaring failure. Son publicly admitted his mistake, citing the core issue as cult-like trust in founder Adam Neumann and a lack of post-investment governance.
  • The 2020 multi-billion dollar loss on Nasdaq call options, dubbed the 'Nasdaq Whale' scandal, was a serious overreach from VC into secondary market derivative speculation, exposing a loss of judgment under COVID-related anxiety.
  • The 2020 deal to sell Arm to Nvidia for $40 billion collapsed due to global antitrust scrutiny and disputes with the Chinese joint venture, wasting two years of strategic time and massive investment banking fees, forcing a shift to an independent IPO path.
  • The Vision Fund's record $32 billion loss in the 2022-2023 fiscal year, with many portfolio companies seeing valuations shrink by over 70% (Katerra and Zume went bankrupt; OYO, Getaround, and Rappi saw massive layoffs), proved that the post-investment capabilities of a mega-fund could not keep up with its capital scale.
  • The long-term losses following the Sprint merger, which resulted in SoftBank's stake being diluted to 24% after the 2020 T-Mobile merger, showed that the dream of a telecom empire failed to deliver expected value, reflecting that SoftBank's integration capabilities for overseas operational assets are far inferior to its eye for internet assets.

关键成功要素

  • Son's core capability is judging the end-game of an industry + the courage to over-bet: He consistently sees the end-game a decade early and has the courage to inject far more capital than peers at stages when others are afraid to bet.
  • SoftBank's 'Fund of Funds + Own Capital' structure allows it to leverage its own capital using money from LPs like Saudi PIF. Of the $93 billion in the Vision Fund, SoftBank only contributed $28 billion but retained control over all investment decisions.
  • Long-term holding of core assets without easy exit: Arm was held through the failed sale and Vision Fund losses from 2016 to its 2023 IPO; Alibaba was held for 22 years from 2000 to 2022. Time is SoftBank's most critical ally.
  • Reconstruction of the AI infrastructure narrative since 2025: Turning the failure of the Vision Fund into the AI infrastructure story of Stargate and OpenAI, using a new grand narrative to reset the capital market's valuation anchor for SoftBank.
  • Possessing a rare global asset network: Arm (chip foundation), T-Mobile/Sprint (telecom), PayPay (payments), OpenAI (models), and Yahoo Japan (portal). These assets create synergy stories that other funds cannot replicate.

Lessons

  • When VC capital scale exceeds the ceiling of post-investment capability, it backfires: The $93 billion Vision Fund's shopping pace far exceeded the number of companies SoftBank could truly perform deep due diligence and post-investment management on. The loss of control in WeWork was not an accident but a structural problem.
  • Trust in a founder's personal charisma cannot replace corporate governance: Son's delegation of power to Adam Neumann was the core trigger for the WeWork disaster. Any large investment must write post-investment governance and key decision-making rights into investment terms, rather than relying on personal relationships.
  • The 'put all eggs in one super basket and ignore it' long-term holding strategy requires an extremely strong balance sheet and debt management capability. SoftBank's forced $41 billion asset sale in 2020 was the reaction to excessive leverage.
  • Secondary market speculation is the easiest trap for VC firms to fall into: The massive loss on Nasdaq call options in 2020 shows that crossing from the primary market into derivative speculation under valuation anxiety is almost guaranteed to fail. Firms should stick to the asset classes they excel at.
  • When the core narrative collapses, have the courage to cut off an arm and reconstruct: SoftBank's 2022-2023 liquidation of Alibaba, exit from telecom operations, and abandonment of the Vision Fund 2 grand narrative to concentrate all chips on Arm and AI was the key to its survival.

Core Data

  • FY2023 Revenue：Approx. 6.76 trillion yen (Public data, not independently verified)
  • Arm IPO Valuation：$54.5 billion (September 2023) (Public data, not independently verified)
  • Arm Acquisition Price：$32 billion (2016) (Public data, not independently verified)
  • Cumulative Investment in OpenAI：Approx. $34.6 billion (holding ~11% stake, end of 2025) (Public data, not independently verified)
  • Stargate Project Scale：$500 billion (with OpenAI, Oracle, MGX) (Public data, not independently verified)
  • Vision Fund Max Annual Loss：$32 billion (FY2023) (Public data, not independently verified)
  • WeWork Loss：Cumulative investment over $10 billion, bankruptcy in 2023 (Public data, not independently verified)
  • Stock Exchange：Tokyo Stock Exchange 9984 (Public data, not independently verified)
  • Founding Year：1981 (Public data)
  • Investment in French AI Data Center：Up to 75 billion euros (announced May 2026) (Public data, not independently verified)
  • Vision Fund Scale：$93 billion (Fund I) + $108 billion (Fund II) (Public data, not independently verified)
  • Alibaba Investment Return：$20 million → $60 billion (~3000x) (Public data, not independently verified)

Competitors / Peers

In the tech VC space, it benchmarks against mega-funds like Tiger Global, Coatue, and D1 Capital, but SoftBank's capital scale and global asset portfolio far exceed these pure financial institutions. In the AI infrastructure space, it benchmarks against Microsoft, Amazon AWS, and Google Cloud; SoftBank has formed an alliance with Oracle through the Stargate project, attempting to occupy positions in data centers, compute, and models simultaneously. In the chip foundation space, it benchmarks against Nvidia, but SoftBank holds architectural dominance in mobile and IoT through Arm, following a path different from GPUs.

Sources

Related

Same industry / scale
Conglomerate / Trading HouseJapan

Panasonic: The Century-Old Electronics Giant Built by a Bicycle Apprentice

Conglomerate / Trading HouseJapan

Toyota Tsusho: From Toyota's Trading Department to a Comprehensive New Energy Trading House in Africa

Conglomerate / Trading HouseJapan

Rakuten, Japan's Largest E-commerce Ecosystem Group

Conglomerate / Trading HouseMulti-region

Samsung SDS: Transitioning from an Internal Group IT Service Provider to a Global Logistics and Cloud Player

⚠️ This entry draws on self-reported accounts and public reporting, not independently verified, and is not investment, startup, or legal advice; revenue figures are case-specific and not representative of typical results.