The Great Liquidation
The honeymoon is over. Masayoshi Son just blinked. On November 10, 2025, SoftBank Group Corp. confirmed a massive divestment of its Nvidia holdings, liquidating approximately 28.5 million shares at an average price of $142.20. The market response was a brutal 10.4 percent sell-off in Tokyo during the November 11 session. This was not a routine rebalancing. This was a tactical retreat that signaled a fundamental schism in the AI investment thesis. For years, SoftBank rode the coattails of Jensen Huang’s hardware dominance. Now, Son is cashing out to chase a specter known as Project Izanagi.
The $14 Billion Tactical Retreat
The scale of this exit is staggering. According to filings processed in the 48 hours leading up to November 11, SoftBank has effectively halved its exposure to the primary driver of the S&P 500’s 2024 and 2025 gains. Analysts at Bloomberg noted that the timing suggests an urgent need for liquidity rather than a lack of faith in Nvidia’s Blackwell architecture. SoftBank’s Loan-to-Value (LTV) ratio had been creeping toward the 25 percent danger zone, and the high-interest environment of late 2025 has made servicing that debt increasingly expensive. By dumping Nvidia, Son has secured a war chest of nearly $14.3 billion, but at the cost of his firm’s credibility as the ultimate ‘AI Bull.’
Technical Breakdown of the Sell-Off
Investors hate uncertainty. When the 13F filings began to leak early on November 10, the panic was immediate. SoftBank’s stock (9984.T) broke below its 200-day moving average for the first time in eighteen months. This technical breakdown was exacerbated by the realization that SoftBank is no longer just an investor in the AI ecosystem but intends to be a direct competitor. The capital from this sale is reportedly earmarked for a $100 billion venture to build AI-specific silicon, a move that pits Son directly against his former golden goose, Nvidia.
Comparison of Recent Institutional Moves
SoftBank is not alone in its caution, but it is the most aggressive in its exit strategy. While other institutional players are trimming at the edges, Son is hacking off entire limbs. The following table illustrates the divergence in strategy as of the latest SEC filings available on November 11.
| Institution | Action Taken (Nov 2025) | Estimated Share Change | Strategy Alignment |
|---|---|---|---|
| SoftBank Group | Major Liquidation | -28.5 Million | Aggressive Pivot to Own Silicon |
| Tiger Global | Minor Trim | -1.2 Million | Risk Mitigation |
| BlackRock | Accumulation | +3.4 Million | Passive ETF Inflow |
| Project Izanagi (Internal) | Capital Injection | +$14.3 Billion | Vertical Integration Venture |
The Izanagi Gamble and the Arm Connection
Why would Son sell the best-performing asset of the decade? The answer lies in Arm Holdings. SoftBank still owns 90 percent of Arm, and the integration of Arm’s architecture into a proprietary SoftBank AI chip is the ultimate goal. Per reports from Reuters on the morning of November 11, Son believes he can replicate Nvidia’s success by controlling the software and hardware stack from the ground up. This is a massive gamble. Nvidia has a ten-year head start in CUDA, the software layer that makes their GPUs indispensable. Son is betting $14 billion that he can bridge that gap in twenty-four months.
The Mechanics of the Drop
The 10.4 percent drop in SoftBank shares reflects a ‘complexity discount.’ Investors bought SoftBank to get indirect exposure to Nvidia and Arm. Now that the Nvidia stake is being cannibalized to fund a speculative R&D project, the valuation model has shifted. The market is pricing in the risk of Project Izanagi failing. If the new chips do not outperform Nvidia’s 2026 roadmap, SoftBank will have traded a liquid gold mine for a concrete bunker. Furthermore, the volume of the sale suggests that SoftBank used several dark pools to execute the trades, which led to a ‘lagged’ realization of the sell pressure once the block trades were reported on the public exchange.
Watching the February 2026 Milestone
The market is now focused on one specific data point. By February 15, 2026, SoftBank is expected to release the first performance benchmarks for the Izanagi prototype silicon. If those numbers do not show a 30 percent efficiency gain over Nvidia’s Rubin architecture, the current sell-off will look like a mild correction compared to what is coming. Investors should ignore the noise about ‘market development’ and focus on the LTV ratio. If it breaches 30 percent before the February reveal, the liquidation of Arm shares will be the next inevitable shoe to drop.
