Why SoftBank Bought 308 Companies in One Day

SoftBank added 308 companies to its consolidated accounts on September 30.

It did not negotiate 308 takeovers, sign 308 cheques or inherit 308 independent operating businesses before lunch.

The Japanese group completed one acquisition: DigitalBridge, a Florida-based alternative asset manager focused on data centres, telecom towers, fibre networks, small cells and edge infrastructure. DigitalBridge arrived with 307 subsidiaries, taking the number of newly consolidated companies to 308.

The number is an accounting consequence of buying a complex global investment platform. It is also a clue to what Masayoshi Son actually purchased for $3.1 billion.

DigitalBridge says it manages $121 billion of assets and has more than 45 portfolio companies. Those assets include some of the world’s better-known digital-infrastructure businesses, such as Vantage Data Centers, Switch, DataBank and Zayo.

SoftBank did not acquire direct ownership of every tower, fibre network or data centre inside that $121 billion universe. It bought the company that manages funds, advises investment vehicles, commits general-partner capital and brings institutional investors into those assets.

The distinction is the story

Son’s AI strategy requires sums that can overwhelm even SoftBank. OpenAI needs enormous quantities of computing capacity. Computing capacity needs chips, power, land, buildings, transmission and fibre. SoftBank has already invested $64.6 billion in OpenAI, raised record amounts of debt and recognised a $3.7 billion loss after revising power-equipment procurement plans for American data-centre projects.

DigitalBridge offers something SoftBank cannot manufacture quickly: a global system for persuading pension funds, sovereign wealth funds, insurers and other institutions to help pay for infrastructure.

SoftBank bought 308 companies in accounting terms. Strategically, it bought an external balance sheet.

Did SoftBank really acquire 308 companies?

SoftBank’s financial report says 308 companies entered the scope of consolidation: DigitalBridge Group and its 307 subsidiaries.

That does not mean each subsidiary is a separately negotiated acquisition or a major operating company. Large alternative-asset managers use layers of legal entities for funds, general partners, advisory businesses, holding structures, regional operations and individual transactions.

SoftBank acquired 100 per cent of DigitalBridge’s common shares and voting rights through a merger. DigitalBridge survived as the operating company and became a SoftBank subsidiary. It continues to be led by chief executive Marc Ganzi as a separately managed platform.

The deal was announced at an enterprise value of approximately $4 billion. SoftBank’s results record a cash payment of $3.1 billion and total accounting consideration of ¥477.4 billion. The difference reflects the distinction between enterprise value, which incorporates debt and other claims, and the cash equity consideration recorded for the acquisition.

The transaction closed on the final day of SoftBank’s reporting period. DigitalBridge’s balance sheet was therefore consolidated at September-end, but the acquisition contributed no meaningful post-acquisition operating period to the six-month profit-and-loss statement.

That means the latest results tell us what SoftBank bought, not yet how the business will perform under its ownership.

Did SoftBank buy $121 billion of infrastructure?

No. Assets under management are not the same as assets owned by the manager.

DigitalBridge’s $121 billion figure includes capital for which it provides investment-management services, along with general-partner capital. The company says the number is generally composed of third-party capital, including commitments that may not yet earn fees, capital not subject to fees or carried interest, and financing used within investments.

The more revealing number is $40.2 billion. That was DigitalBridge’s fee-earning equity under management at June 30, according to its regulatory filing.

In other words, the headline AUM figure was approximately three times its fee-earning equity base. That does not make the $121 billion claim invalid. It shows why AUM must not be mistaken for revenue-producing capital or property belonging to SoftBank.

DigitalBridge reported $175.1 million of fee revenue in the first half of 2026, almost unchanged from the corresponding period a year earlier. Its total revenue moved far more dramatically because carried-interest allocations and principal-investment income can fluctuate with asset performance and realisations.

SoftBank has therefore bought an asset manager, not a passive vault of infrastructure. Its economics will depend on management fees, performance-related income, investment gains and the ability to raise increasingly large pools of capital.

Why does SoftBank need other people’s money?

Because AI infrastructure is moving from billions to tens and potentially hundreds of billions of dollars.

SoftBank-backed SB Energy plans an eight-gigawatt IT campus in Ohio under 20-year leases to OpenAI. The first approximately 800 megawatts alone are expected to require an investment of $30 billion to $40 billion. SoftBank is also developing projects in Texas, France and Malaysia.

Financing those projects entirely through SoftBank equity or corporate borrowing would place enormous pressure on the group’s balance sheet. Consolidated interest-bearing debt increased by approximately ¥7.9 trillion during the six months to September, while finance costs more than doubled.

SoftBank says it intends to use project finance for AI power and data centres, ideally on a non-recourse basis. Under that structure, lenders are repaid primarily from a project’s own cash flows and have limited recourse to the parent. SoftBank’s risk can, in principle, be capped at the equity it contributes.

DigitalBridge can add a second source of capital: institutional investors participating through managed funds and investment vehicles.

The potential model is straightforward. SoftBank can act as an anchor investor or strategic partner. DigitalBridge can originate a project, conduct due diligence, raise third-party equity and manage the asset. Banks and credit investors can provide project debt. Long-term customer contracts can support the financing.

That structure allows SoftBank to pursue infrastructure larger than the amount it could safely fund alone. It also gives DigitalBridge access to projects generated by SoftBank’s relationships with OpenAI, Arm, Nvidia and energy developers.

Is DigitalBridge SoftBank’s new infrastructure bank?

It is closer to an investment manager and capital-raising engine than a bank.

DigitalBridge brings relationships with public and private pension funds, sovereign wealth funds, insurance companies, endowments and other asset managers. These institutions are attracted to infrastructure because long-term contracts can produce durable cash flows that behave differently from venture-capital returns.

SoftBank’s original Vision Fund model concentrated on equity investments in technology companies. DigitalBridge offers access to real assets and fund-management fees. Data centres and fibre networks can be financed, operated and eventually sold or refinanced in ways that a minority holding in an unlisted startup cannot.

That could make SoftBank’s AI exposure more balanced. It could also produce a recurring fee business less dependent on quarterly movements in portfolio valuations.

The acquisition price indicates that SoftBank expects considerable future earning power. It recorded provisional goodwill of ¥276.3 billion against total consideration of ¥477.4 billion. Goodwill therefore represented approximately 58 per cent of the consideration at closing.

SoftBank says the goodwill reflects the expected earning power of DigitalBridge’s future business and synergies with existing operations. The purchase-price allocation remains provisional and can change within a year as acquired assets and liabilities are identified and measured.

That accounting treatment reveals the bet. A large part of what SoftBank paid for cannot be touched like a server rack or measured like a tower. It consists of people, fundraising relationships, deal pipelines, operating expertise and the expectation that the combined companies can create more business together.

Can SoftBank simply use DigitalBridge’s investors to fund its own projects?

No. DigitalBridge is not a captive cheque book.

It manages money on behalf of outside institutions and owes duties to those investors. A SoftBank-backed data centre must still meet the return, risk and governance standards of the relevant fund before DigitalBridge can commit client capital.

Ganzi has described DigitalBridge as SoftBank’s third-party infrastructure arm but has also stressed that investments must satisfy independent return requirements. That separation is essential. Pension funds and sovereign investors will not provide capital merely to reduce pressure on SoftBank’s balance sheet.

Conflicts must also be managed. SoftBank may want infrastructure built rapidly to support OpenAI and its wider AI ecosystem. DigitalBridge’s clients may prefer stronger contractual protection, lower construction risk or higher returns. The manager cannot automatically place a SoftBank priority above the interests of its funds.

DigitalBridge’s continued operation as a separately managed platform is therefore not cosmetic. It is central to retaining investor confidence.

What could derail SoftBank’s plan?

The first danger is that AI infrastructure demand may be genuine while individual projects remain unfinanceable.

Data-centre developments require power, transmission, land, permits, water solutions, equipment and creditworthy customers. A signed proposal or announced gigawatt target does not guarantee construction. Ganzi has warned that a large share of newly proposed projects may fail to proceed because of zoning, regulation or financing constraints.

The second danger is price. Heavy competition for data-centre assets can compress future returns. If infrastructure valuations are already elevated, DigitalBridge must raise more capital without overpaying for the assets meant to generate fees and carried interest.

The third danger is the quality of the AUM. Only part of the $121 billion headline currently produces management fees, and some commitments may never be deployed. Growth in reported AUM is valuable only if capital becomes fee earning and ultimately delivers acceptable returns.

The fourth danger is integration without control. SoftBank wants strategic alignment, while DigitalBridge must retain enough independence to attract outside capital. Too little coordination weakens the promised synergies. Too much turns the manager into a financing arm for its owner and could frighten limited partners.

Why buy the manager instead of creating one?

Time. SoftBank could hire infrastructure specialists and attempt to raise its own third-party funds. Building a record, winning institutional mandates and creating an international operating network would take years.

DigitalBridge arrives with more than 100 digital-infrastructure professionals, over 45 portfolio companies and three decades of collective industry experience. Its funds already manage capital across data centres, towers, fibre, small cells and edge facilities.

The acquisition gives SoftBank immediate entry into the club that controls infrastructure capital. It can bring projects and strategic relationships to the table while DigitalBridge brings investors, underwriting expertise and operating partners.

The 307 subsidiaries are evidence of that machinery. They reflect a business constructed to move money across funds, jurisdictions and assets. The complexity is not automatically a virtue, but it is difficult to reproduce quickly.

What did Masayoshi Son really buy?

He bought a way to make SoftBank’s ambitions larger than SoftBank’s own capital.

OpenAI supplies the demand narrative. Arm supplies a critical layer of computing technology. SB Energy develops power and data-centre capacity. DigitalBridge is supposed to bring institutional money and infrastructure discipline into the system.

The strategy makes sense precisely because SoftBank’s latest results show its limits. The group raised $56.5 billion in one quarter, increased its debt and still absorbed a $3.7 billion loss after changing power-equipment plans. Building every project directly on its own balance sheet would magnify that strain.

DigitalBridge offers a possible escape from that constraint. It can distribute the capital requirement across funds, lenders and partners while SoftBank retains strategic influence and participates in the upside.

The challenge is that other people’s money comes with other people’s rules.

SoftBank acquired one company on September 30. Accounting turned it into 308. The valuable number, however, is not 308 or even $121 billion. It is the amount of outside capital DigitalBridge can persuade to follow Masayoshi Son into the AI infrastructure race.