Softbank Group shares fall as OpenAI exposure overshadow bumper Q4

SoftBank reported a 1.829 trillion yen net profit for the March quarter, far above the 295.2 billion yen estimate and more than tripling year over year, but shares fell nearly 4% as investors focused on rising exposure to OpenAI. The company’s OpenAI stake was valued at $79.6 billion, implying a $45 billion cumulative gain, but concerns are building around the $17.5 billion remaining balance on a $40 billion bridge loan and the startup’s governance, spending, and competitive position. S&P Global Ratings had already downgraded SoftBank’s outlook in March, citing weaker liquidity and financial capacity tied to the OpenAI bet.

Analysis

This is less a clean earnings beat than a balance-sheet stress test for the AI capex cycle. SoftBank’s mark-to-market gains are increasingly tethered to a single private asset whose valuation is being validated by fundraising rounds, not cash flow, so the equity is behaving like a levered call option on late-stage AI sentiment rather than a diversified conglomerate. That makes the setup fragile: if private-market multiples compress, the equity can re-rate faster than the reported P&L can normalize because the debt stack introduces a liquidity penalty that public investors will price in before any realized loss appears.

The second-order winner is Google, not just as a model competitor but as a “good enough, cheaper, and already monetized” alternative that can pressure OpenAI’s pricing power and funding narrative. If enterprise buyers conclude frontier performance is converging, the market may rotate from speculative AI beneficiary names toward hyperscalers with self-funded capex and operating cash flow. ARM is a cleaner beneficiary than SoftBank itself only if AI infrastructure spend keeps broadening beyond one model provider; otherwise its demand is still hostage to the same capex enthusiasm that is now under scrutiny.

Near term, the key risk is not an earnings miss but a sentiment air pocket in the next 4-8 weeks if OpenAI headlines turn from growth to governance, usage, or IPO skepticism. Over 6-12 months, the bigger catalyst is financing: any widening in credit spreads or tightening in private-markets liquidity would force investors to discount SoftBank’s asset value at a larger haircut, especially given the pledged-asset structure. The contrarian angle is that the market may be overestimating how much of SoftBank’s NAV is actually impaired today; as long as OpenAI can keep raising at higher marks, the equity can stay elevated despite the leverage, but that path dependency is the whole risk.

Key Decisions for Investors

• Short 9984 on a 1-3 month horizon into any post-earnings bounce; target a move toward the stock’s debt-adjusted NAV discount if OpenAI sentiment de-risks. Risk/reward is attractive because downside can extend quickly if the market re-prices the private valuation stack.
• Pair trade: long GOOGL / short 9984 for 2-4 months. Thesis: Google offers AI exposure with balance-sheet self-funding and lower governance risk, while SoftBank is a levered proxy for a single private valuation. Use tight stop-loss if OpenAI announces another up-round or IPO path.
• Add ARM only on broad AI semis weakness, not on SoftBank strength. ARM benefits if infrastructure spend broadens across multiple model vendors; otherwise it is exposed to the same capex unwind. Prefer a staged entry over 4-6 weeks and hedge with QQQ puts.
• Sell upside calls on 9984 into volatility spikes over the next 30-60 days. The leverage to private-markets sentiment makes upside discontinuous, but implied vol should stay elevated; monetize premium while keeping directional exposure limited.
• If you want direct AI upside, rotate from 9984 into hyperscaler baskets over the next quarter. The better risk/reward is in firms that can absorb AI spend from cash flow, not those depending on asset marks and bridge financing.