SoftBank’s scramble in late 2025 to fund another $22.5 billion investment in OpenAI turned out to be only one stage of a much larger financial commitment. By February 2026, the Japanese investment group had agreed to put another $30 billion into OpenAI, potentially taking its cumulative investment in the artificial-intelligence company to $64.6 billion and its ownership to approximately 13%.
Financing that strategy has required SoftBank to use several parts of its balance sheet. The company has sold major equity holdings, expanded borrowing against assets, arranged a $40 billion bridge facility and continuously managed its loan-to-value ratio as CEO Masayoshi Son concentrates more capital around artificial intelligence. SoftBank’s September decision to repay $25.9 billion of outstanding bridge loans early shows that the financing strategy is evolving almost as quickly as the investments themselves.
The result is a very different story from the original question of whether SoftBank could find $22.5 billion. It did. The more important question now is how much financial concentration the company is willing to accept as OpenAI becomes one of the most important assets in SoftBank’s portfolio.
The Original $22.5 Billion Commitment Was Part of a $40 Billion Deal
In March 2025, OpenAI announced a $40 billion funding round led by SoftBank at a $300 billion post-money valuation. SoftBank initially funded $10 billion and intended to provide additional capital later in the year, subject to conditions related to OpenAI’s corporate restructuring.
SoftBank also had the ability to syndicate part of the investment to other investors. The structure meant its eventual direct funding requirement was lower than the headline $40 billion figure but still extraordinarily large for a single private-company investment.
By late 2025, attention centered on the remaining approximately $22.5 billion SoftBank needed to provide. Reports indicated that the company was preparing to sell assets and increase borrowing to meet that obligation, raising questions about how aggressively Son was reshaping the group’s portfolio around AI.
SoftBank ultimately completed the investment. By the end of its 2025 fiscal year, the company said its cumulative investment in OpenAI had reached $34.6 billion.
SoftBank Then Added Another $30 Billion
Rather than slowing down after completing the 2025 transaction, SoftBank doubled down.
On February 27, 2026, the company entered into a definitive agreement for an additional $30 billion investment in OpenAI through SoftBank Vision Fund 2. SoftBank said completing that investment would take its cumulative OpenAI investment to $64.6 billion and give it approximately 13% ownership.
The new commitment was structured as three $10 billion tranches scheduled for April, July and October 2026. SoftBank completed the first tranche on April 1 and its annual reporting says another $10 billion was funded in July, leaving the final $10 billion scheduled for October.
The size of the commitment means OpenAI is no longer simply another large Vision Fund investment. It has become central to SoftBank’s strategy for capturing value from the growth of artificial intelligence.
That creates enormous upside if OpenAI continues increasing in value, but it also increases SoftBank’s exposure to the performance of one private technology company.
A $40 Billion Bridge Facility Helped Finance the Expansion
SoftBank’s financing became more aggressive in March 2026 when it arranged a $40 billion bridge facility, primarily to fund the additional OpenAI investment.
The facility was provided by JPMorgan Chase, Goldman Sachs, Mizuho Bank, Sumitomo Mitsui Banking Corporation and MUFG Bank. It was scheduled to mature in March 2027 and was unsecured, according to SoftBank’s disclosures.
SoftBank borrowed $10 billion from the facility in April to fund the first 2026 OpenAI tranche. The company subsequently drew additional funds as it executed the second tranche and managed other financing requirements.
Bridge loans are designed to provide capital before longer-term financing becomes available. SoftBank said from the beginning that it expected to repay the borrowings through existing assets and other financing measures rather than leave the entire facility outstanding until maturity.
That distinction is important because the $40 billion facility did not represent permanent financing for OpenAI. It gave SoftBank flexibility to execute its investment commitments first and determine the optimal combination of asset monetization and longer-term financing afterward.
SoftBank Is Already Repaying the Bridge Debt
The financing story changed again in September.
On September 9, SoftBank announced that it would repay the entire $25.9 billion outstanding balance under the bridge facility on September 15, well ahead of the March 2027 maturity date. SoftBank said it had drawn a cumulative $30 billion from the facility by that point.
The early repayment demonstrates why focusing only on the gross amount of new borrowing can produce a misleading picture of SoftBank’s finances. The group frequently moves between asset sales, secured financing, bridge loans and other instruments as it manages investments and liquidity.
That does not eliminate the financial risk associated with its AI strategy. It does show that the company retains substantial ability to monetize assets and restructure financing when required.
SoftBank’s financial strategy therefore needs to be evaluated through both sides of its balance sheet: the debt it assumes and the enormous portfolio of technology assets against which that financing is managed.
Asset Sales Have Been Central to the Strategy
SoftBank’s willingness to sell valuable holdings has been one of the clearest ways it has generated liquidity for its AI investments.
The company reduced its T-Mobile US position, generating billions of dollars that could be redirected toward new investments. Reuters reported in February that SoftBank had sold approximately $12.7 billion of T-Mobile shares as it increased its exposure to OpenAI and other AI-related assets.
This follows a familiar pattern under Masayoshi Son. SoftBank has repeatedly treated major equity holdings as sources of capital that can be monetized when it identifies what Son considers a more important strategic opportunity.
The trade-off is concentration. Selling diversified public holdings to finance a much larger position in a private AI company can increase dependence on the success of that company and the broader AI investment cycle.
SoftBank is effectively reallocating parts of its existing technology portfolio toward what it believes will become the dominant technological platform of the next era.
Arm Is Another Important Source of Financial Flexibility
SoftBank’s ownership of Arm is central to its balance-sheet strength. The British chip-design company has become one of SoftBank’s most valuable holdings as investor enthusiasm around artificial intelligence has increased demand for semiconductor-related assets.
SoftBank has used Arm shares to support financing arrangements, allowing it to raise capital without necessarily selling the underlying stake. Reuters reported that SoftBank expanded a margin loan backed by Arm shares to $20 billion as part of its broader financing strategy.
Asset-backed borrowing can be attractive because it allows SoftBank to retain economic exposure to an appreciating asset while obtaining immediate liquidity.
But it also introduces market risk. If the value of collateral falls sharply, leverage can become more burdensome and lenders may require additional protection. SoftBank therefore closely monitors the relationship between debt and the value of its holdings.
That relationship is captured in one of the company’s most important financial metrics: loan-to-value.
SoftBank Watches Its Loan-to-Value Ratio Closely
SoftBank measures its financial discipline partly through its LTV ratio, calculated using adjusted net interest-bearing debt relative to the equity value of its holdings.
The company says its policy is to maintain LTV below 25% under normal conditions, with an upper threshold of 35% even during emergencies. That framework is designed to prevent investment activity from creating leverage that becomes unsustainable if technology valuations fall.
SoftBank’s LTV had risen to approximately 20.6% by the end of December 2025, according to Reuters, as the company increased borrowing and committed more capital to OpenAI. That remained within SoftBank’s stated normal limit but illustrated the balance-sheet consequences of the investment campaign.
The company’s own 2026 annual report makes clear that management views LTV as one of the variables it can directly control. If market conditions threaten to push leverage higher, SoftBank says it can monetize holdings, use asset-backed financing or pursue hybrid financing.
The strategy depends on SoftBank continuing to hold assets that remain sufficiently valuable and liquid to provide those options.
OpenAI Has Already Produced Large Paper Gains for SoftBank
The scale of the investment becomes easier to understand when viewed alongside the gains SoftBank has already recognized from OpenAI.
SoftBank reported a $19.8 billion investment gain related to OpenAI for the October–December 2025 quarter, helping the company post its fourth consecutive quarterly profit. The gain reflected increases in the estimated value of SoftBank’s OpenAI holdings rather than equivalent cash received from selling the position.
That distinction is important. Private-company valuation gains can significantly increase reported investment income, but they do not provide the same liquidity as selling publicly traded shares.
OpenAI’s valuation nevertheless rose dramatically as investor demand for frontier AI companies increased. Its February 2026 financing valued the company at approximately $840 billion, substantially above the valuation associated with SoftBank’s original 2025 investment.
For SoftBank, that appreciation supports Son’s argument that aggressive investment can create enormous portfolio value. It also increases the financial consequences if OpenAI’s valuation later falls.
OpenAI Is Becoming a Concentration Risk
SoftBank itself now explicitly identifies its OpenAI exposure as an important risk.
In its 2026 risk disclosures, the company notes that it had invested $34.6 billion in OpenAI as of March 31 and subsequently committed another $30 billion. It also highlights the uncertainty surrounding AI because of rapid technological development, changing competitive conditions and evolving regulation.
That is a significant change from the diversified Vision Fund model associated with SoftBank’s earlier investment era. The group still owns a broad portfolio, but OpenAI represents an increasingly large single-company commitment.
The risk is not simply that OpenAI could fail. Its valuation could grow more slowly than expected, competitors could gain market share, AI economics could become less attractive or infrastructure costs could place pressure on profitability.
OpenAI itself requires extraordinary amounts of capital to train models and build computing infrastructure. The company is expected to consume enormous quantities of cash as it expands, meaning investors may need to continue providing financing even as valuations rise.
SoftBank Is Building a Wider AI Portfolio Around OpenAI
SoftBank’s strategy is broader than purchasing OpenAI equity.
The group is attempting to position itself across several layers of the AI economy, including semiconductor intellectual property through Arm, AI infrastructure, data centers, robotics and large-scale computing projects.
SoftBank is also pursuing its acquisition of DigitalBridge, the digital-infrastructure investment manager with exposure to data centers, fiber and related assets. That transaction fits the same thesis: artificial intelligence requires not only models but also enormous quantities of physical infrastructure.
Gignomist’s analysis of SoftBank’s DigitalBridge acquisition examines how that deal could give the group another platform for financing AI infrastructure at scale.
The common thread is Son’s belief that AI will become foundational infrastructure across the global economy. SoftBank is attempting to gain exposure to several parts of that value chain rather than betting only on one model developer.
OpenAI, however, remains the largest and most visible component of the strategy.
The AI Infrastructure Bill Keeps Getting Bigger
The amount of capital required to compete in frontier AI is increasing rapidly. Training advanced models requires large clusters of specialized processors, while serving hundreds of millions of users creates enormous ongoing inference costs.
That means OpenAI’s financing needs do not end when SoftBank completes its next $10 billion payment. Reuters reported this week that OpenAI could burn through roughly $278 billion in cash between 2026 and 2030, based on Financial Times reporting.
Those projections can change substantially as revenue, model efficiency and infrastructure strategy evolve. They nevertheless illustrate why investors are discussing AI financing in amounts that would have appeared extraordinary for a software company only a few years ago.
The capital intensity also changes the role of investors such as SoftBank. Providing tens of billions of dollars is no longer simply a way to purchase shares in an AI company; it helps finance the computing infrastructure required for the company to continue developing and operating its models.
That connects SoftBank’s investment strategy directly with the wider race to finance chips, electricity and data centers.
The Final 2026 OpenAI Tranche Is Still Ahead
SoftBank’s latest financial disclosures say $20 billion of its new $30 billion OpenAI commitment had been funded by July, through two $10 billion tranches.
The remaining $10 billion is scheduled for October 2026, although SoftBank notes that closing dates could be accelerated if OpenAI shares are publicly listed.
Once that commitment is completed, SoftBank says its cumulative investment in OpenAI should reach $64.6 billion.
That number puts the original $22.5 billion financing challenge into perspective. What once looked like an extraordinary final payment has become one stage in a capital commitment almost three times as large.
SoftBank’s challenge has therefore shifted from finding enough money to complete one OpenAI transaction toward maintaining financial flexibility while continuing to finance a much broader AI strategy.
SoftBank Is Repeating Its Highest-Conviction Investment Model
Masayoshi Son has built SoftBank around unusually concentrated technology bets before. The company’s early investment in Alibaba became one of the most successful technology investments ever made, generating enormous value that later helped finance other parts of SoftBank’s portfolio.
The Vision Fund era produced a much more mixed record. Investments in companies including WeWork demonstrated the risks of combining abundant capital with aggressive growth assumptions and high private-market valuations.
OpenAI represents another high-conviction phase, but on a larger scale than most individual Vision Fund bets.
Son is effectively arguing that artificial intelligence represents a technological transition important enough to justify reallocating huge amounts of SoftBank’s capital around it. If that thesis proves correct, OpenAI and the surrounding AI infrastructure portfolio could reshape SoftBank’s value.
If valuations decline or the economics of frontier AI disappoint, the same concentration could amplify losses.
The $22.5 Billion Story Has Become a Balance-Sheet Story
When SoftBank was preparing to fund its remaining OpenAI commitment in 2025, the immediate question was whether it could raise the required $22.5 billion through asset sales and borrowing.
By September 2026, that question has largely been answered. SoftBank completed its 2025 investment, committed another $30 billion, funded two-thirds of that new commitment and arranged tens of billions of dollars of temporary financing. It has now also moved to repay $25.9 billion of bridge borrowing ahead of schedule.
The more durable question is what this financing strategy says about SoftBank itself.
The company is increasingly using its portfolio as a financial engine for AI investment. Public holdings can be sold, valuable stakes can support borrowing, bridge facilities can fund transactions quickly and debt can later be refinanced or repaid as SoftBank reorganizes its balance sheet.
That provides enormous financial flexibility, but it also ties SoftBank’s future more closely to the performance of AI assets.
The original $22.5 billion payment was therefore not the climax of SoftBank’s OpenAI investment story. It was evidence of a larger transformation already underway: Masayoshi Son is rebuilding SoftBank around the belief that artificial intelligence will create enough value to justify one of the most aggressive capital-allocation strategies in global technology.
