The consensus on SoftBank Group (9984) right now is straightforward: Arm is the engine, the yen is the tailwind, and the investment portfolio is finally working. Three-month momentum on the Nikkei 225 has run from the low 52,000s to nearly 59,513 per index data, and 9984 has been riding that wave hard enough to sit at turnover rank number one on the TSE with a 3.93% single-day gain per kabutan.jp. The narrative writes itself — recovery, momentum, re-rating. I’ve read this chapter before.
Here’s what that story skips: free cash flow came in at negative ¥1.4279 trillion for FY2025.03 per kabutan.jp, against an operating cash flow of just ¥203.5 billion in the same period. Think of SoftBank’s income statement as a beautifully renovated house — stunning on the outside, but the plumbing hasn’t been updated in years. The cumulative Q3 FY2026.03 net profit of ¥3.1726 trillion per kabutan.jp is genuinely eye-catching, but a large portion of that figure is mark-to-market appreciation, not cash hitting the bank account. When the market celebrates a profit surge of that magnitude without asking where the cash is, I get nervous in the way I only get nervous when I’ve seen this before.
The yen depreciation angle is real, and I won’t dismiss it. A weaker yen translates the dollar-denominated value of Arm Holdings and other Vision Fund positions into larger yen figures almost automatically, which is why earnings visibility looks so clean right now. The TSE governance reform push has also been a legitimate structural tailwind, with foreign inflows into large-cap Japanese equities compressing the risk premium on names like this one. These are real forces. What they miss is the other side of the currency trade: SoftBank runs significant global debt obligations, and yen weakness raises the effective cost of servicing any unhedged foreign-currency liabilities. The same exchange rate that inflates the asset column quietly pressures the liability column. Currency tailwinds feel permanent until they don’t.
Copper at $6.068 per Yahoo Finance commodity data is worth a moment’s attention here. Copper tracks industrial and data center construction activity globally, and at this level it supports the thesis that demand for high-performance computing infrastructure — the market Arm’s IP licensing feeds directly — remains structurally intact. I don’t want to oversell one commodity reading, but when the underlying hardware cycle is healthy, Arm’s royalty and licensing revenue has a real floor under it. That’s the one part of the bull case I find genuinely durable.
Let me interrogate my own assumptions for a moment, because the contrarian reflex can become its own trap. The cumulative Q3 FY2026.03 EPS of ¥553.5 per kabutan.jp, stacked against the FY2025.03 full-year EPS of ¥195.2 per kabutan.jp, suggests a real acceleration in earnings power — not just asset inflation. If Arm’s semiconductor cycle exposure continues to compound and the group executes well on strategic divestments from legacy non-core holdings, the operating cash flow picture could normalize faster than I’m giving it credit for. The FY2026 full-year earnings release on May 13, 2026 per kabutan.jp will be the first clean read on whether that normalization is actually happening or whether it’s still a mark-to-market story dressed in operating clothes.
The valuation is where the consensus feels most comfortable and where I feel least so. At ¥5,424 per kabutan.jp, the stock is trading fractionally above the analyst consensus average target of ¥5,375.61 per Yahoo Finance, with the high end of the range at ¥7,570 and the low at ¥3,100. A PBR of 1.98x per kabutan.jp is not outrageous for a holding company with Arm at its core, and the ROE of 10.15% per kabutan.jp gives the book value some credibility. The trailing P/E is literally incalculable right now per kabutan.jp — earnings have been volatile enough that no clean trailing multiple exists — and that kind of ambiguity usually gets resolved in the direction of the cycle, not against it. If the semiconductor cycle softens and Vision Fund marks come in, there’s nothing in the trailing multiple to anchor the downside.
If the May earnings release shows operating cash flow inflecting meaningfully upward and free cash flow moving toward breakeven, my skepticism is wrong and the consensus re-rating story has real legs. That’s the invalidation condition, and I’m watching for it honestly. Until then, the stock is essentially priced at fair value on the best-case scenario — which means the market is already charging you for the optimism before it’s been delivered.
I’ve watched enough investment holding companies go through cycles to know that the gap between reported profit and actual cash generation is where the surprises live — and they are rarely pleasant ones.
© The Nonexpert · Original