9984 enters the post-earnings week having pulled off one of the more striking vindication trades of the year.
The stock has risen 58% over the past month and 10.4% in the last five days alone. The catalyst is now public: Q4 EPS of $1.02, up from $0.29 a year earlier, with a $45 billion gain on its OpenAI position driving annual profit to $32 billion. Today also brought a separate headline — Arm and SoftBank reportedly made a last-ditch approach to acquire AI-chip startup Cerebras weeks before its expected IPO. That bid was rebuffed, but it underlines how aggressively Masayoshi Son is pressing the AI theme.
The most striking thing about the lending market is how completely it has de-risked ahead of the print. Short Interest as a percentage of free float has collapsed from 1.85% in early April to just 0.44% now — a drop of more than 60% in six weeks. That is not a slow bleed-off; it reads like a concerted cover. Availability in the borrow market reflects this: with so few shares short, supply of borrows well exceeds demand, and cost to borrow remains loose at 0.83% annualised — cheap by almost any standard, and roughly in line with where it spent most of April. The ORTEX short score of 26.5 (ranked in the 87th percentile for low short pressure) confirms the picture: whoever was leaning against the AI thesis through early April has largely stepped aside.
The Street's read is cautious relative to where the stock now trades. The consensus mean price target is around ¥5,483 — fractionally below the current price of ¥5,987. The P/E has expanded sharply, now running at 74.6x, up more than ten points over the past thirty days. Price-to-book has moved similarly, rising 0.7x over the same period to 1.87x. The EV/EBITDA of 26.5x drifted up roughly 0.5x over the month. None of these are cheap multiples for a Japanese holding company, and the dividend yield is negligible — the dividend history in the data is stale (last recorded payment dates to 2022), so income is not the buy thesis here. The factor scoring shows some of the tension: the EV/EBIT score ranks in the sixth percentile, flagging stretched valuation, while the dividend score of 84 is something of an artefact given that payment record.
Ownership reinforces how concentrated this story remains. Masayoshi Son holds 33.96% of shares, and his last reported change was a reduction of 48 million shares — though that was filed in November 2025. BlackRock added just over 2.2 million shares through April, and Capital Research added 2.4 million. The marginal institutional buying looks more like index-related drift than conviction accumulation. The headline is that Son's concentrated grip and the stock's AI leverage mean this is not a diversified holding for most institutions.
Recent earnings reactions have been painful on first read. The May 8 print delivered a one-day move of -10.6%, and the February 12 release produced -6.7% the next day with a further -5.6% over the following week. Those reactions came into results, not out of them — suggesting the stock had been well-positioned into the prints and sold the fact. The dynamic this week looks different: the sharp short cover and re-rating into the announcement suggest positioning had already turned constructive before the numbers landed. Next scheduled results are June 24.
With shorts near a multi-month low and the stock trading through consensus targets, the question going into June is whether the Cerebras approach, the OpenAI mark-up, and Son's broader AI deployment narrative can sustain a premium above where analysts currently have their numbers.
See the live data behind this article on ORTEX.
Open 9984 on ORTEX →ORTEX Market Intelligence content is generated by AI from a snapshot of ORTEX's proprietary data. Content is informational only and does not constitute investment advice.
