The Fund That Bet the House on the Future

In the spring of 2017, a Japanese telecom entrepreneur named Masayoshi Son walked into the technology industry with a checkbook so large it changed the physics of the room. His new vehicle, run out of a townhouse on Grosvenor Street in London, closed with roughly $98.6 billion in committed capital - more money in a single fund than the entire American venture-capital industry had raised in many prior years. The firm that manages it is called SoftBank Investment Advisers, and its job, then and now, is to decide where that mountain of money goes.

Legally it is SB Investment Advisers (UK) Limited. In practice it is the engine room of the SoftBank Vision Funds, the largest technology-focused investment funds ever assembled. Since 2017 the firm has backed 332 companies, produced 129 unicorns and logged 69 exits. Its portfolio reads like a map of the modern internet: Arm and Nvidia in the silicon, Coupang and DoorDash and Uber and Grab in the apps, ByteDance in the feed, and, most consequentially of late, OpenAI in the models. Few investors touch that many layers of the same industry at once. Fewer still do it on purpose.

01What it actually does

Strip away the scale and SoftBank Investment Advisers does something familiar: it raises money from large investors, buys equity stakes in fast-growing technology companies, and tries to return more than it put in. What makes it unusual is the size of each move. Where a traditional venture firm might write a $10 million seed check and hope to own a sliver, the Vision Fund has written checks large enough to reshape a company's entire trajectory - and, sometimes, its market. The thesis has always been blunt: back enormous ambition with enough capital that competitors cannot keep pace.

The capital comes from a short list of very large partners. SoftBank Group, the Japanese conglomerate Son built, is the anchor. For the first fund, Saudi Arabia's Public Investment Fund committed roughly $45 billion and Abu Dhabi's Mubadala about $15 billion, with strategic corporates including Apple, Foxconn, Qualcomm and Sharp rounding out the roster. The second fund, launched in 2019 with about $56 billion in committed capital, leaned far more heavily on SoftBank itself.

The business model underneath is a standard one dressed in unusual numbers. As the manager of the funds, SoftBank Investment Advisers earns fees for running the capital and a share of the gains it generates, and it aims to realize those gains over a medium-to-long horizon through IPOs, secondary sales and other exits. What separates it from a typical fund is the holding period and the concentration: it is built to sit with large positions for years and to keep supporting them through multiple rounds rather than rotating quickly in and out. The firm calls the belief behind this "the transformative power of technology" - the wager that a handful of companies, given enough runway, will grow into the infrastructure everyone else depends on.

02The full-stack bet

The clearest way to understand the firm today is to look at how it frames its own portfolio. Rather than sorting investments by stage or geography, it organizes them by layer of the AI stack - hardware at the bottom, infrastructure in the middle, applications on top. The idea is that value in artificial intelligence does not sit in one place; it moves up and down a stack, and a fund large enough can own pieces at every level.

This is where SoftBank Investment Advisers differs from most of its peers. A specialist chip fund would never also own a food-delivery app; a consumer-internet investor would rarely wade into semiconductor design. The Vision Fund's structural advantage - and its structural risk - is that it does both, and treats the connections between them as part of the thesis. Arm designs the processors in most of the world's smartphones; Uber, DoorDash and Coupang are businesses that only exist because those phones are everywhere.

03Who it serves

The firm has two sets of customers, and they sit on opposite sides of the balance sheet. On one side are the limited partners - SoftBank Group, the sovereign wealth funds, the strategic corporates - whose money it is entrusted to grow. On the other are the founders and companies who receive that money and, with it, access to follow-on financing, hiring help, introductions and a path toward an eventual IPO. A Vision Fund check has historically come with more than cash: it comes with the implicit promise that the fund can keep writing checks as a company scales.

Beyond the check, the firm has built expertise that is hard to buy off the shelf. Operating from London, Silicon Valley and Tokyo, its investment teams specialize in late-stage, cross-border deals - the kind that require understanding both a Korean e-commerce market and an American logistics one, both a semiconductor roadmap and a consumer app's unit economics. It also turns its own portfolio into a research asset: through work it publishes under the "Sozo Pulse" banner, the firm gathers proprietary data from across its companies - surveying operators such as chief human-resources officers on how they are adopting AI - and feeds those signals back into how it invests.

04The problems it solves

For the fastest-growing technology companies, the constraint is often not ideas but capital - specifically, patient capital available in amounts large enough to fund years of expansion before profits arrive. That is the gap SoftBank Investment Advisers was built to fill. By concentrating capital rather than spreading it thin, it lets a small number of companies pursue strategies that would otherwise be impossible to finance: building global logistics networks, subsidizing user growth across continents, or funding the multi-year, capital-hungry work of designing chips and training frontier models.

That same concentration cuts both ways, and the firm has never hidden from it. Its most public misfires - most famously a large bet on the office-space company WeWork - became case studies in what happens when conviction outruns discipline. The lesson visibly reshaped how the firm operates.

05How the money has moved

The returns have been anything but steady. After the turbulence of the WeWork era and a broad technology sell-off, the funds swung back sharply as the AI trade took hold. A single position - OpenAI - now sits at the center of the story.

Figures are approximate, drawn from SoftBank's reported results; relative bar lengths are illustrative.

06The 2025 turn

In September 2025 the firm did something a fund built on diversification rarely does: it narrowed. The Vision Fund cut roughly 20% of its workforce and redirected capital away from a sprawling startup portfolio toward large-scale AI and chip infrastructure, including a major position in OpenAI and participation in data-center and compute initiatives. The message was that the next phase would not be about owning hundreds of small bets, but a handful of very large ones at the base of the AI stack.

The leadership changed to match. Rajeev Misra, the strategist who helped Son raise more than $150 billion and build the Vision Fund from scratch, stepped back from his co-CEO role in 2024 - notably, just as the funds were posting their best gains in years. Alex Clavel took the lead, working alongside investment leaders including Vikas Parekh and Sumer Juneja and a team of more than 50 professionals across London, Silicon Valley and Tokyo.

07Where it sits in the market

There is no true peer at the same scale. Growth investors such as Tiger Global, Coatue, Thrive Capital and General Catalyst compete for the same late-stage deals, and firms like Andreessen Horowitz and Sequoia shape the same technology cycles. But the sovereign and strategic pools that once fed the Vision Fund - Saudi Arabia's PIF, Abu Dhabi's Mubadala, Singapore's Temasek and GIC - increasingly invest directly, which means SoftBank Investment Advisers now competes with some of its own former backers for the biggest positions.

What has not changed is the underlying wager. From the first $98.6 billion fund to the AI-infrastructure pivot of 2025, SoftBank Investment Advisers has operated on the belief that technology compounds faster than caution - and that the investor willing to be wrong at the largest scale is also positioned to be right at the largest scale. Whether the OpenAI-anchored era vindicates that view or repeats the volatility of the past is, for now, an open question. The firm is not hedging it.