Masayoshi Son Has 67% of SoftBank's U.S. Stock Portfolio in Intel. He Didn't Buy a Single Share Last Quarter.

SoftBank’s latest 13F shows it held 86,956,522 Intel shares at both March 31 and June 30, with the position’s value rising from $3.8B to $12.1B purely due to Intel’s stock nearly tripling. By the end of June this made Intel 67% of SoftBank’s disclosed U.S. holdings, but the article argues no new conviction was added; as Intel repriced, the stake is about $8.4B (~$3.7B less than the filing). Despite improving fundamentals—Q2 revenue +25% YoY to $16.1B and AI/data-center revenue +59% YoY—the stock is described as still expensive at ~58x adjusted earnings, implying limited near-term support from the 13F.

Analysis

The market is likely mispricing the signal in the 13F: this is not a fresh sponsor bid, it is a mark-to-market story layered on top of Intel’s own equity issuance. That matters because the marginal buyer of the stock was the company itself financing capex, which improves liquidity but also tells you equity holders are still being asked to fund the turnaround. In that setup, upside is increasingly a function of execution beating very high expectations, while downside is amplified by dilution and a crowded “recovery” narrative.

The more interesting second-order effect is competitive: if Intel’s capital plan works, the winners are not SoftBank or legacy holders but the ecosystem that can monetize Intel’s investment cycle without relying on Intel’s foundry economics. TSMC and AMD are the cleaner beneficiaries of any delay in Intel’s self-sufficiency, while semiconductor equipment names benefit only if third-party foundry demand actually materializes; if the foundry remains mostly internal, the capex simply slows the dilution of the equity story, not the strategic gap.

Near term, the stock can still trade on momentum and AI positioning, but the next 1-3 months should force a re-underwriting around margins, cash flow, and external foundry traction. The contrarian miss is that “big holder” optics do not create a floor when the stake is static and the company has just expanded share count at nearly the same price. What would invalidate the bearish setup is a sustained beat on gross margin/FCF plus evidence that foundry revenue is becoming third-party-led rather than internal load-balancing.

Key Decisions for Investors

• Short INTC into strength or buy a 1-3 month put spread around current levels; the trade is against multiple compression once the market focuses on dilution and the absence of incremental sponsor buying. Falsify if the stock reclaims and holds above the post-offering high on materially better guidance.
• Prefer a relative-value long TSM / short INTC pair for the next 1-3 months; the market is paying Intel a premium for a turnaround that still depends on unproven foundry economics, while TSM’s cash-flow visibility is cleaner. Cover if Intel shows meaningful external foundry bookings or capex discipline without growth sacrifice.
• Use SOXX as a hedge rather than an outright directional short if you want to isolate Intel-specific risk; this avoids getting run over by broader AI/semi momentum. Add only if INTC fails to hold the $95-100 area after the next catalyst.
• Do not treat SFTBY as a clean long on the basis of this stake; the Intel mark-up is not incremental buying, so there is limited evidence of active support. Any SFTBY exposure should be driven by its broader portfolio, not by presumed Intel sponsorship.