SoftBank is not a stock so much as a leveraged opinion about the future of technology, and its journal entry opens with numbers that would be typing errors anywhere else. Up 227.6% over the year to our June cycle. A two-year total return of 267%. A typical daily move of 7.7%, the most violent character in our entire coverage, with a 50.8% drawdown in the same record that produced a 1.35 risk-adjusted score. The framework reads Markup at ¥6,472, which is simultaneously a tripled stock and 29% below its 52-week high of ¥9,074. Three days ago the quantitative state model dropped out of its bull classification into Sideways. On a tape this fast, that is not a footnote; it is the first paragraph.

The Investor Read: What Season Is This Stock In?

The season is high summer with storm sirens. The Markup label is honest: even 29% below its high, the structure of higher lows that carried the tripling has not broken, and the 96% annualised return over the data window is the strongest in our Japanese coverage. But every other line in the table tells you what kind of summer this is. The trailing multiple of 7.4 is an accounting artefact of investment gains, and the forward 21.9 is the market’s estimate of what the operating reality costs. The screen fails twice, on a 42.0% debt ratio and on revenue purity at 8.0%, which together describe the structure exactly: a heavily levered holding company whose income is mostly not the thing its name suggests. Our conservative fair-value model prints ¥13,312, a 47.6% margin of safety, double the price, and for once it is more bullish than the street’s ¥7,000 median. When our harshest model is the optimist, the honest conclusion is that nobody’s model prices this thing well; it is a bet on its portfolio, and the journal will score the label, not the theology.

The Trader Read: What Does the Tape Look Like Now?

Tactically, respect the character number above everything: 7.7% typical daily movement means this instrument does a normal stock’s monthly range before lunch, and sizing, not opinion, decides who survives it. The map: the June marker at ¥6,461 is where the tape has parked for three weeks, a 12.5% up-month at that reading having stalled; the ¥9,074 high is 40% away; and the state model’s three-day-old exit from its bull classification is the single most timely fact on this page. Fresh state exits on violent tapes resolve fast: either the pause becomes a base and the bull state returns, or the 50.8% drawdown history reminds everyone it was not decoration. There is no boring outcome here. The tactical read updates in the daily sessions.

The Tension: Everything About This Entry Is the Tension

Most entries in this journal have one fact arguing against the thesis. SoftBank’s file is arguments all the way down, so we log the two strongest on each side. For the bulls: a two-year record of 267% with a 1.35 risk-adjusted score that most funds would frame, and our own conservative model at double the price, the only name in the batch where our floor-of-pessimism number sits far above the market. For the bears: a state model that just pulled its endorsement after the run of a lifetime, and a screen fail that is not cosmetic, because 42% debt against a portfolio of volatile technology stakes is how 50.8% drawdowns happen and how the next one gets bigger. The insider and political files are empty, logged as empty. For the values-based investor the double fail settles it structurally: this name stays outside the screened universe regardless of the label. For everyone else, this page now exists, and its dates will do the arguing.

What Would Change the Read

• The state model: the bull classification returning within weeks logs the exit as a pause; the sideways reading decaying towards bear on this tape is the fire alarm, and we will date it within days.
• The June shelf: weekly closes under the ¥6,461 area would put the first crack in the markup structure that survived the 29% pullback.
• The portfolio cycle: this stock is a derivative of its technology stakes. The same dispersion we logged across the AI chain applies here at one remove, and a turn in that cycle re-prices this file overnight.
• The debt line: deleveraging that brings the 42.0% ratio under our threshold would remove one of the two screen fails; the revenue-purity fail is structural to the holding-company form.