SoftBank loses $50b in market cap as AI stock slump deepens

Shares of SoftBank Group fell 6.9% on November 7, extending a weekly drop of nearly 20% amid a broader selloff in AI-related stocks.

SoftBank, which has major investments in AI and owns a controlling stake in UK-based semiconductor designer Arm Holdings, saw its market capitalization shrink by almost US$50 billion this week.

Arm shares listed in the US also declined 1.2% overnight.

David Gibson of MST Financial said the drop in SoftBank shares reflected investor caution about high valuations in the AI sector and uncertainty over potential partnerships involving OpenAI.

SoftBank reportedly considered acquiring US chipmaker Marvell Technology earlier this year, according to Bloomberg.

Other Japanese tech stocks, including Advantest, Renesas Electronics, and Tokyo Electron, also fell.

🔗 Source: CNBC

🧠 Food for thought

Implications, context, and why it matters.

SoftBank NAV discount narrows to 40% in Q1

NAV reached ¥32.4 trillion in Q1 2026, while the NAV discount tightened to 39.8–40%. The market had started to price in SoftBank’s AI exposure before this week’s drop 1.
The near 20% weekly slide tracked sector-wide AI valuation jitters that outweighed company factors. Arm’s Q2 revenue guide of $1.01–1.11 billion implies 19.7–31.5% year-over-year growth, and the loan-to-value ratio sits at 17% 2 1.
SoftBank tripled its Nvidia stake to $3 billion in Q1 2025 and owns 90% of Arm Holdings. Arm recently topped $120 billion in value, yet the stock traded around a 40% NAV discount in Q1 2026. The selloff may have opened a pricing gap without signaling weaker business trends 3 1.

Tech teams can use loosened GPU supply chains as the AI selloff pressures hyperscalers and cloud providers

AI infrastructure teams and startups could see easier GPU buys as speculative demand cools. Margin pressure on cloud providers may open multi-year deals for Nvidia H100 or H200 data center GPU capacity with specialized GPU clouds (smaller providers focused on renting GPU compute) or hyperscalers (massive cloud platforms such as Amazon Web Services, Microsoft Azure, and Google Cloud) at better prices 4.
Enterprise AI buyers should watch for any changes in Nvidia data center GPU allocation or lease pricing. The stock fell with SoftBank, which may push adjustments that open procurement windows that were closed during the peak AI hype cycle 2.