The AI-focused equity rally lost momentum this week despite Taiwan Semiconductor Manufacturing Co. reporting a strong second-quarter earnings beat [1].

This trend highlights a growing tension in the semiconductor market. While the physical production of chips remains in high demand, investors are becoming increasingly sensitive to forward-looking guidance from the companies that design them.

TSMC reported that its second-quarter revenue increased 36% to NT$1.27 trillion [1]. Following the announcement on July 13, the company's share price rose 1.7% on the Taipei exchange [1]. The results suggest that the fundamental demand for AI hardware continues to scale rapidly.

However, the broader market reaction has been muted due to conflicting signals from other industry leaders. Earlier this month, Broadcom saw its stock decline 14% following its own earnings report [2]. Although Broadcom beat some earnings expectations, it provided a soft outlook for AI chips, a move that dragged down other sector players including AMD and Intel [2].

Broadcom's AI-chip revenue guidance was set at $16 billion, which missed the $17.2 billion estimate projected by analysts [2]. This discrepancy has created a contradiction in market sentiment. Some analysts argue that TSMC's revenue beat proves demand is still surging, while others suggest Broadcom's miss signals a cooling period for the sector [1], [2].

Despite these fluctuations, some executives remain optimistic about the long-term trajectory of the industry. Hock Tan said that AI semiconductor revenue guidance is exceeding $100 billion into 2027 [2].

The current volatility reflects a shift in investor behavior. The market is moving away from rewarding general AI growth and is instead scrutinizing specific revenue targets, and delivery timelines for next-generation chips.

“TSMC reported that its second-quarter revenue increased 36% to NT$1.27 trillion”

The divergence between TSMC's growth and Broadcom's missed guidance indicates a decoupling between semiconductor manufacturing and chip design. While the infrastructure for AI continues to be built at a record pace, the market is now questioning whether the end-users of these chips can sustain the current rate of spending, leading to a more cautious valuation of AI-related stocks.