Brian Nichols is the co-founder of Angel Squad, a community where you’ll learn how to angel invest and get a chance to invest as little as $1k into Hustle Fund's top performing early-stage startups

In 1995, Masayoshi Son met Jerry Yang and David Filo, two Stanford PhD students who were building a website directory they called Yahoo. Son spent six minutes with them and wrote a check for $2 million. Yahoo went public and Son's stake was eventually worth over $3 billion. Five years later, he invested $20 million in a Chinese e-commerce company called Alibaba at a $60 million valuation.

That stake would grow to over $130 billion by 2018. Masayoshi Son investments are defined by one characteristic that no other investor on the planet has matched: the willingness to write enormous checks into companies and categories that the rest of the market has barely imagined yet. The upside has been extraordinary. The downside has been equally dramatic. Understanding both is one of the most useful exercises in venture capital education.

The SoftBank Origin and the Yahoo Bet

Son founded SoftBank in 1981 at age 24, initially as a software distributor in Japan. He built it into one of Japan's largest telecom companies while simultaneously investing in technology companies in the US and China at a pace that confused most observers at the time. By the late 1990s, SoftBank had invested in hundreds of internet companies, and Son's paper net worth briefly exceeded $75 billion during the dotcom boom, making him the wealthiest person in the world for a few weeks in early 2000. The bubble burst, and SoftBank's portfolio value collapsed by over 90%.

Son rebuilt. He acquired Vodafone Japan in 2006 and transformed SoftBank into Japan's third-largest wireless carrier. He negotiated the acquisition of Arm Holdings for $32 billion in 2016, a bet on the semiconductor architecture underpinning everything from mobile devices to data centers. Arm went public on Nasdaq in September 2023, and SoftBank's 90% stake is now worth more than SoftBank's entire market capitalization.

The Vision Fund Era

In 2017, Son raised the SoftBank Vision Fund, a $98.6 billion vehicle backed primarily by Saudi Arabia's Public Investment Fund ($45 billion) and SoftBank itself ($28 billion). It was the largest technology-focused investment fund in history by a factor of several. The thesis was to identify AI-driven companies across every sector and write checks at a scale that would let them dominate their categories.

The first fund backed Uber, DoorDash, Coupang, DiDi, Grab, Oyo, Fanatics, and WeWork, among others. The WeWork implosion in 2019 became the defining disaster of the era. SoftBank ultimately invested over $10 billion in the coworking company, and when WeWork's IPO plans collapsed and the company filed for bankruptcy in November 2023, the losses were catastrophic. Vision Fund also infamously sold its entire Nvidia position in early 2019 for a $3.3 billion gain, only to watch those shares grow to over $150 billion in value by mid-2024. Son publicly called it "the fish that got away."

By 2025, Vision Fund had sold down or written off $29 billion in US-listed holdings, including Coupang, DoorDash, and Grab. Vision Fund 2, originally targeted at $108 billion, struggled to attract external capital and is largely funded by SoftBank itself.

The AI Pivot and Stargate

Since 2023, Son has been repositioning SoftBank as an AI infrastructure company rather than a pure venture investor. By the end of 2025, SoftBank had invested approximately $34.6 billion in OpenAI, becoming its third-largest shareholder after the OpenAI nonprofit foundation and Microsoft. In January 2025, Son was named chairman of the Stargate Project, a joint venture among SoftBank, OpenAI, Oracle, and MGX, backed by $500 billion in committed AI infrastructure investment over four years. The project aims to build data centers and AI compute capacity across the United States.

Elizabeth Yin, Hustle Fund GP, has noted that the best investors identify platform shifts before the market prices them in. Son identified AI as the dominant platform shift of the next century and realigned every significant SoftBank capital decision around it, selling Alibaba (fully divested by mid-2024), T-Mobile US, and Vision Fund positions to fund the pivot. In October 2025, SoftBank also agreed to acquire ABB's robotics division for $5.375 billion, describing it as an entry into "Physical AI."

Eric Bahn, Hustle Fund GP, has pointed out that the scale at which Son operates makes his decisions less instructive for early-stage investors than for anyone thinking about macro capital allocation. But one principle does translate: the willingness to make a decision and act at the size of your conviction is rare, and Son has consistently done it whether it produced Alibaba returns or WeWork losses.

Angel Squad and the Conviction Sizing Lesson

Masayoshi Son investments carry a lesson that applies at every check size: the quality of a bet is largely determined before the capital is committed, not after. Son's best returns came from his fastest decisions, the Yahoo check after six minutes, the Alibaba bet in 2000. 

His worst losses came from following conviction with additional capital into situations that weren't working. Angel Squad trains investors to develop pre-investment conviction frameworks that hold up under scrutiny, not just enthusiasm. With 2,500 members across 50 countries, the community gives investors a sounding board for exactly this kind of thesis development before they deploy capital. Visit hustlefund.vc/squad to learn more.

The Takeaway

Masayoshi Son has produced the highest single-investment return in venture history with Alibaba, built and largely collapsed the world's largest startup fund, and is now positioning himself as the architect of America's AI infrastructure decade.

The scale of the swings is instructive in itself. At any level of investing, the willingness to think in terms of transformative platform shifts rather than incremental opportunities is what separates extraordinary returns from average ones. The conviction has to be earned first, though. That's the part Son's worst bets demonstrate most clearly.