SoftBank Leads Asian Tech Selloff Amid Rising AI Infrastructure Costs

SoftBank Group shares fell on June 26, 2026, triggering a broad selloff in Asian technology stocks as investors raised concerns over the rising costs of artificial intelligence infrastructure. According to market reports, the decline in SoftBank’s valuation sparked a ripple effect across semiconductor and software firms in Tokyo, Seoul, and Taipei.

The selloff impacted several high-weight technology entities. Market data shows declines for Arm Holdings PLC, which is heavily owned by SoftBank, as well as Taiwan Semiconductor Manufacturing Co Ltd (TSMC) and Hon Hai Precision Industry Co Ltd. In China, Tencent Holdings Ltd, Baidu Inc, and Xiaomi Corp also saw share price drops during the session.

Why did SoftBank and AI stocks fall on June 26?

The downturn stems from growing skepticism regarding the capital expenditure required to sustain AI growth. Investors are questioning whether the massive spending on data centers and specialized hardware will yield immediate financial returns. This shift in sentiment hit SoftBank Group particularly hard due to its aggressive investments in the AI ecosystem.

The volatility extended to the semiconductor equipment sector. ASML Holding NV, Infineon Technologies AG, and ASM International NV experienced downward pressure. These companies provide the machinery necessary to produce the chips that power AI, making them sensitive to any perceived slowdown in infrastructure investment.

How did the selloff impact Asian markets?

The selloff began in Japan and quickly spread to other regional hubs. SoftBank Corp and Advantest Corp both saw declines in Tokyo. The contagion reached Taiwan, where TSMC—the primary manufacturer for most global AI chips—faced selling pressure. Hon Hai Precision Industry Co Ltd, a key assembler of AI servers, followed a similar trajectory.

Chinese tech firms also retreated. Baidu Inc and Tencent Holdings Ltd, both of which have integrated generative AI into their core search and social products, saw their valuations dip. Xiaomi Corp, which has been expanding its AI-integrated hardware, also recorded losses.

The broader market impact is visible in the performance of sector-specific trackers. The iShares Semiconductor ETF and the SPDR S&P Semiconductors index both reflected the downward trend, indicating that the selloff was not limited to a few individual companies but was a systemic move away from AI-exposed assets.

What is the connection to US tech giants?

The turmoil in Asia mirrors concerns regarding the “Magnificent Seven” in the United States. Companies like Microsoft Corp, Alphabet Class A, and Meta Platforms Inc are the primary buyers of the infrastructure provided by TSMC and NVIDIA. When these US giants signal that AI costs are mounting without a proportional increase in revenue, it creates a negative feedback loop for the Asian suppliers.

Apple Inc and Qualcomm Inc also remain tied to this cycle. Their reliance on advanced semiconductor nodes means any reduction in overall AI infrastructure spending could lead to lower order volumes for the foundries and designers in Asia.

Market analysts have noted a contrast in how this selloff is framed compared to previous corrections. While earlier dips were often tied to interest rate hikes, the June 26 decline is specifically linked to the “cost of build-out.” This suggests a transition from speculative excitement about AI capabilities to a pragmatic evaluation of the balance sheet.

Which semiconductor firms are most exposed?

The exposure is highest for firms providing the “picks and shovels” of the AI gold rush. This includes:

• TSMC: As the sole manufacturer for high-end AI accelerators.
• ASML: As the provider of EUV lithography machines required for the smallest chips.
• Micron Technology Inc: Which provides the high-bandwidth memory (HBM) essential for AI processing.
• STMicroelectronics NV and BE Semiconductor Industries NV: Which support the broader power and packaging infrastructure.

The synchronized drop across these entities suggests that the market is pricing in a potential plateau in the aggressive infrastructure spending phase of the AI cycle.