SoftBank Group released its quarterly earnings on November 9, with market attention focused on three areas: its dense bets across the artificial intelligence (AI) supply chain, its steadily swelling interest-bearing debt, and the ripple effects of Chinese regulators’ rejection of its acquisition of ABB’s robotics business.
According to Tech in Asia, as of the end of June this year, SoftBank’s net interest-bearing debt had climbed to 10.8 trillion yen (about $68.2 billion), up sharply from 6.6 trillion yen (about $41.7 billion) a year earlier. Meanwhile, SoftBank’s share price fell 11% on November 6 and another 4.2% on November 9, bringing its year-to-date decline to 23%. Investors are assessing whether weakness in Arm’s share price could undermine SoftBank’s ability to borrow against its holdings—SoftBank currently has a roughly $25 billion loan tied to Arm.
ABB Robotics Deal Stalled; China Business May Face Carve-Out
On November 5, ABB said in a statement on its website that SoftBank had been informed that the office of China’s foreign investment security review mechanism had prohibited SoftBank from acquiring ABB’s robotics business in China. That puts the $5.375 billion deal, a year in the making, on hold.
According to Bloomberg, SoftBank and ABB are studying a carve-out of the China business to preserve the possibility of acquiring the operations in other regions. China is the largest single market for ABB’s robotics business, accounting for about 30% of its global operations. ABB robotics products operate widely in Chinese factories, and ABB has robotics production and R&D bases in China; regulators’ concerns about leaks of core data are believed to be an important reason the review was not approved.
From a financial standpoint, the deal was originally quite attractive for SoftBank. ABB’s disclosed figures show that in 2025 its robotics business generated $2.331 billion in revenue and $174 million in net profit after tax, implying a valuation of about 30 times earnings based on the purchase price. By comparison, Midea Group’s 2017 acquisition of Kuka carried a price-to-earnings multiple of about 50 times.
Full AI Supply Chain Takes Shape as Debt Pressure Mounts
The ABB Robotics acquisition is not an isolated move but a key step in SoftBank’s effort to build a full AI supply chain. A review of its investment track record in recent years shows a fairly clear blueprint:
DateTargetAmountSegmentJuly 2024Graphcore (UK AI chip design)UndisclosedUpstream chipsAugust 2025Foxconn’s Lordstown plant and equipment in OhioUndisclosedMidstream computing infrastructureNovember 2025Ampere Computing (AI chip design)$6.5 billionUpstream chipsSeptember 2026DigitalBridge (digital infrastructure investment management)$3.1 billionMidstream computing infrastructure—13% stake in OpenAI$64.6 billionDownstream large models—ABB robotics business (blocked)$5.375 billionDownstream physical AI
Note: SoftBank’s investment amount in OpenAI comes from a Tech in Asia report; the other transaction amounts come from public disclosures.
Upstream, SoftBank is moving into AI chip design through Graphcore and Ampere Computing; midstream, it is using the Foxconn plant and DigitalBridge to build out hardware manufacturing and computing infrastructure; downstream, it holds OpenAI, a core large-model asset, and is seeking to complete its physical AI and embodied intelligence capabilities at the device end through ABB Robotics. Its Yahoo Japan unit owns internet assets including a news portal, search, local services and LINE, providing natural venues for application deployment.
Mirroring its “buy, buy, buy” approach is asset shuffling. In the fiscal year ended March 2025, SoftBank exited its investment in DoorDash; in the fiscal year ending March 2026, it has sold a cumulative 75.4 million shares of U.S. telecom operator T-Mobile US, raising $16.25 billion. In October 2025, SoftBank also fully liquidated its stake in Nvidia, a move the market interpreted as a monetization to raise cash for a bigger bet on OpenAI.
Stargate and Regulatory Variables
After the U.S. midterm elections, the Democrats won both chambers of Congress and will lead the legislative and appropriations agenda of the 120th Congress starting in 2027. House Democratic Leader Jeffries has already put data centers on the AI regulatory agenda. The most closely watched U.S. data center project, Stargate, involves SoftBank, Oracle and OpenAI, with plans to invest $500 billion over four years and an initial phase of $100 billion, with SoftBank leading the finances and fundraising.
Some lawmakers have proposed legislation to pause new construction and expansion of AI data centers. If such restrictions take effect, Stargate’s not-yet-started campuses and subsequent expansions could be affected. But for cash-strapped SoftBank, a project slowdown may not be entirely bad news—after consuming large amounts of capital for its OpenAI equity investment, SoftBank has little room left to keep funding Stargate, and short-term funding pressure could actually ease as a result.
The real risk lies in the pace of OpenAI’s IPO. SoftBank’s most pressing need now is to push OpenAI to list as soon as possible, but tighter regulation or a weaker market could delay the listing further, forcing SoftBank to bear debt-servicing pressure for longer.
Market Impact
Amid a cooling in global AI investment in the third quarter, SoftBank’s share price has pulled back deeply since the fourth quarter, falling more than 40% in a single quarter and erasing all of its year-to-date gains. Beyond secondary-market volatility, the core issue is this: before its “chips—computing infrastructure—algorithmic models—terminal robots” map is fully connected and its portfolio companies can generate their own cash flow, SoftBank must prioritize stable cash flow and smooth debt repayment.
For investors, SoftBank has become a high-beta play on the AI theme, and whether its strategy pays off depends on whether the company can use its existing asset-shuffling tools to ride out the current debt-repayment cycle and period of external regulatory adjustment. China’s regulatory rejection of the ABB deal adds new uncertainty to its physical AI ambitions.