SoftBank's Vision Fund: how a $93bn fund reshaped technology investing

Executive overview

SoftBank is not a bank. It is a Japanese conglomerate built by Masayoshi Son — a serial entrepreneur who turned a software distribution business into one of the world's most powerful investment vehicles.

The Vision Fund is the largest fund ever raised in any asset class, at $93bn (targeting $100bn), with a 12-year time horizon. Its creation was made possible by SoftBank's $3.3bn acquisition of alternative asset manager Fortress Investment Group — a deal that gave SoftBank the infrastructure, compliance machinery, and guaranteed management fee streams needed to operate at this scale.

The core insight: SoftBank structured the Vision Fund as much for its predictable $18bn+ in management fees over 12 years as for the investment returns — turning venture capital into a cashflow business.

Masayoshi Son: background and early bets

• Born 1957 in Kyushu, Japan, to Korean immigrant parents; family forced to take a Japanese surname
• At 16, cold-called the president of McDonald's Japan, flew to Tokyo uninvited, and secured a meeting; was told to learn English and study computers
• Moved to the US within weeks; graduated UC Berkeley (economics and computer science) having already sold an electronic translator to Sharp for $1.7m and made ~$1.5m importing Space Invaders machines
• Returned to Japan in 1980, changed his name back to Son, and founded SoftBank as a software distributor
• Expanded into PC publishing (magazines), then acquired Ziff Davis and Comdex to dominate US tech media and events in the mid-1990s
• Launched a US venture capital arm; Brad Feld (later of Foundry Group) was one of the early partners

The internet boom, the crash, and the pivot

• SoftBank invested in ~800 companies in the late 1990s; key wins included Yahoo (became largest shareholder at IPO) and Yahoo Japan (a joint venture with Jerry Yang)
• At the peak of the bubble, Son briefly became the world's wealthiest person; SoftBank's market cap reached ~$200bn
• When the bubble burst, Son lost an estimated $70bn in personal wealth — the largest personal wealth loss in history at the time
• Lesson drawn: cashflow and profits matter; pivoted SoftBank toward infrastructure businesses with predictable revenue
• Invested in Japanese broadband, acquired Japan Telecom, then identified mobile as the next platform shift

The iPhone deal and SoftBank Mobile

• Around 2005–06, Son flew to the US with a hand-drawn sketch of an iPod with a phone — essentially the iPhone concept — and pitched Steve Jobs
• Jobs laughed but agreed: if Apple built such a device, SoftBank could be the exclusive Japan carrier
• Son returned to Japan and raised ~$20bn in debt financing (largely via Deutsche Bank's Rajiv Misra) to acquire Vodafone Japan
• In 2008, SoftBank Mobile became the exclusive iPhone carrier in Japan; the company tripled its market share
• Later acquired Sprint in the US (majority stake); also bought ARM Holdings (UK chip designer) for $32bn in 2016 — completing stakes across mobile's value chain: commerce (Alibaba), carriers (SoftBank Mobile, Sprint), and chip IP (ARM)

The Alibaba investment: possibly the best investment ever made

• In 2000, Son invested $20m in a then-obscure Chinese e-commerce company called Alibaba
• Alibaba IPO'd in 2014 — 14 years later — in what was then the largest IPO ever
• SoftBank's stake was worth ~$60bn at IPO: a return too large to calculate as a multiple
• This single outcome, combined with Yahoo and Yahoo Japan gains, generated the capital base that seeded the Vision Fund

Building the Vision Fund

• 2016: Son met Saudi Arabia's deputy crown prince in Tokyo; in a 45-minute meeting secured a $45bn commitment from the Public Investment Fund of Saudi Arabia
• SoftBank committed $25bn of its own capital; total initial close: $70bn
• First close in May 2017 at $93bn, with additional investors including Mubadala (UAE), Apple ($1bn), Qualcomm ($3bn), Sharp, Foxconn, and Larry Ellison's family office
• Fund term: 12 years plus a 2-year extension — longer than any comparable vehicle
• Target: 70–100 investments at an average ~$1bn per deal
• Rajiv Misra (formerly Deutsche Bank, UBS, and briefly Fortress) named to run the fund, based in London

The Fortress acquisition: why it mattered

• February 2017: SoftBank announced acquisition of Fortress Investment Group for $3.3bn — a ~40% premium on a publicly traded firm
• Fortress managed ~$70bn across private equity, hedge funds, and credit funds — zero technology focus
• The acquisition looked baffling: a tech-oriented investor buying a second-tier Wall Street credit manager
• The real logic: Fortress provided the operational infrastructure (1,000+ staff, compliance, trading desks, fund administration) that SoftBank lacked entirely
• Rajiv Misra had briefly worked at Fortress before joining SoftBank — a direct connection
• Fortress was also the first large PE/hedge fund firm to go public (2007), pioneering the management-company-as-equity structure

The management fee logic

• Typical VC/PE fee structure: 2% annual management fee on committed capital, plus 20% of profits (carried interest)
• The Vision Fund's non-SoftBank capital (~$75bn) generates an estimated $18bn in management fees over the fund's 12-year life — essentially guaranteed regardless of investment performance
• SoftBank's $25bn commitment draws no management fee but earns 100% of the profits on that tranche
• Even if the Vision Fund lost all its invested capital, the fee streams would be highly profitable
• This is the infrastructure play: SoftBank structured a business line (SoftBank Financial Services, run by Misra from London) that is cashflow-positive by design

SoftBank Financial Services: the bigger picture

• March 2018: SoftBank announced a new division — SoftBank Financial Services — housing both the Vision Fund and the retained Fortress funds (~$40bn after divesting the fixed-income book)
• Combined AUM: ~$140bn, already the world's second-largest fund manager behind KKR (~$168bn, built over decades)
• Stated goal: double AUM in five years
• Fortress is likely not the last acquisition; SoftBank has signalled it will buy more asset management firms
• The model mirrors Masa's post-bubble infrastructure thesis: own the cashflow-generating layer, then use those cashflows to fund the next bet

The Vision Fund's investment thesis

• The stated vision: own significant stakes in companies underpinning global shifts driven by AI — across transportation, food, work, medicine, and finance
• Portfolio at time of episode: Uber (15% stake, ~$8bn investment), WeWork, DoorDash, Wag, Slack, Flipkart ($4bn)
• SoftBank takes large, sometimes controlling stakes — more influence than typical venture investors
• Unique structural advantage: billion-dollar checks that can move in unified fashion (e.g., the Uber tender offer) — impossible to coordinate across many separate LPs
• Invests in cash-flow-negative, high-growth companies — venture mindset, private equity scale
• Long time horizon means portfolio companies can stay private through the growth phase rather than being forced to IPO or get acquired

Risks and criticisms

• When you raise and deploy large capital into competing companies, customer acquisition spend wars benefit Google and Facebook ad platforms, not the companies themselves
• Capital abundance trains organisations to spend irrationally; they may never learn lean unit economics (contrast: Zappos learned discipline during the dot-com bust)
• Valuation inflation: Vision Fund-scale capital competes with itself across ride-sharing, food delivery, and other sectors, sustaining losses at scale
• Governance misalignment: management companies taken public (as Fortress was) divorce fee income from the investors whose capital generates those fees
• Retail investors have no direct access to Vision Fund returns — though SoftBank itself is publicly traded, providing one indirect route

Grading the Fortress acquisition

• Ben's grade: A — Fortress was essential infrastructure for creating a new financial product that no one else could have built; the fee economics alone justify the $3.3bn price
• David's grade: A− — agrees on the strategic logic but flags SoftBank as a potentially dangerous and controlling counterparty for portfolio companies; long-run execution risk remains
• Both hosts note: paying a 40% premium for a non-technology firm looked irrational in early 2017; in hindsight it was the key structural enabler of a $140bn+ asset management business built in under two years