
SoftBank shares jumped 16.5% as Japan's Nikkei 225 hit record highs. The rally reflects a catch-up trade in AI-linked chip stocks and datacenter infrastructure.
SoftBank Group shares climbed 16.5% on Thursday, leading a broad-based surge in Japanese technology equities as the Nikkei 225 index reached record highs. The move follows an extended holiday period in Japan, forcing local markets to reconcile with a significant global rally in artificial intelligence-linked assets that occurred while Tokyo was closed. This catch-up trade has re-established the correlation between Japanese semiconductor suppliers and U.S. tech benchmarks, particularly the Nasdaq Composite.
The primary driver of Thursday’s price action is the delayed reaction to three sessions of global risk-on sentiment. During the Japanese Golden Week holiday, U.S. markets saw aggressive buying in AI-exposed names, with the Nasdaq Composite hitting record levels. Global X ETFs investment strategist Billy Leung noted that the Nikkei 225 is effectively pricing in three days of global market movement in a single session. This volatility is concentrated in the semiconductor supply chain, where liquidity remains highest for investors seeking exposure to the AI infrastructure build-out.
For those tracking the sector, the performance of ATEYY (Advantest Corp) and Tokyo Electron serves as the most direct proxy for this trend. Advantest shares rose 7.8% while Tokyo Electron gained 9.2%, reflecting their status as essential equipment providers for the global chip manufacturing cycle. Renesas Electronics also participated in the rally with a 13.8% gain, underscoring the breadth of the move across the Japanese tech complex.
SoftBank’s 16.5% gain, its strongest single-day performance since 2020, is tied to its concentrated exposure to the semiconductor and AI software ecosystem. The firm functions as a listed proxy for ARM (Arm Holdings plc) and OpenAI. The market is currently pricing in the potential for these holdings to capture value from the shift toward agentic AI systems and inference-heavy workloads. As these systems scale, the demand for underlying infrastructure has become the focal point for institutional positioning.
Market sentiment is currently supported by the read-through from AMD (Advanced Micro Devices Inc.), which recently reported quarterly results that highlighted the critical role of CPUs in the AI stack. While much of the initial AI investment focused on GPUs, the industry is increasingly recognizing that CPUs are essential for managing agent sandboxes, orchestration servers, and database layers.
According to Rolf Bulk of The Futurum Group, the total addressable market for datacenter CPUs is projected to reach $120 billion by 2030, representing a compound annual growth rate of over 35%. This secular demand narrative provides a floor for the current rally, as investors look past short-term volatility to the long-term capital expenditure cycle in data centers. The following table outlines the performance of key players during this session:
The sustainability of this rally depends on the continued stability of global risk assets and the absence of further geopolitical shocks. Easing tensions between the U.S. and Iran, evidenced by a decline in oil prices, provided a secondary tailwind for Thursday's session. However, the reliance on high-beta tech names means that any reversal in U.S. semiconductor performance—particularly if AMD or ARM face valuation pressure—will likely lead to immediate liquidity outflows in Tokyo.
AlphaScala’s current data reflects a moderate sentiment profile for these names, with ATEYY holding an Alpha Score of 72, while ARM and AMD sit at 62 and 59 respectively. Investors should monitor whether the current price levels hold as the market moves past the initial catch-up phase. A failure to maintain these gains would suggest that the rally was purely a technical adjustment to the holiday gap rather than a shift in fundamental valuation. For deeper market analysis on these trends, monitor the ongoing capital expenditure cycles in the semiconductor space.
AI-drafted from named sources and checked against AlphaScala publishing rules before release. Direct quotes must match source text, low-information tables are removed, and thinner or higher-risk stories can be held for manual review.