Chip Stocks End Seven-Day Decline Amid AI Infrastructure Scrutiny

Chip stocks recently concluded their longest losing streak since 2008. The Philadelphia Stock Exchange Semiconductor Index (SOX) registered a seven-day decline, a downturn not seen since November 2008. This period of underperformance reduced the SOX index by 11%.

Several prominent chip manufacturers experienced notable corrections during this period. Nvidia, a key player in AI acceleration hardware, saw its stock price fall by 13% over three trading sessions. Taiwan Semiconductor Manufacturing Company (TSMC), a critical foundry for advanced chips, also recorded a three-day loss of 10%. Broadcom, another significant semiconductor firm, declined by 11% across two trading days. These movements underscore investor apprehension within the sector.

The market’s scepticism intensified despite a recent rally. Earlier in June, Nvidia briefly became the world’s most valuable company. This surge was predicated on high demand for its graphics processing units (GPUs), essential components for artificial intelligence development and deployment. However, the subsequent correction suggests a re-evaluation of long-term growth trajectories and the capital expenditures required to sustain them.

Analysts have begun to question the immediate and sustained profitability of current AI infrastructure investments. While companies are committing substantial capital to AI capabilities, the rate at which these investments convert into tangible returns remains a subject of debate. Data centre spending, driven by AI, is projected to reach approximately $200 billion in 2024. This represents a significant increase, yet the operational leverage and ultimate profitability from these expenditures are not universally clear.

The SOX index had demonstrated robust growth before this recent downturn, appreciating 38% for the year and 70% over the preceding 12 months. This substantial prior growth provides context for the current correction, indicating that some rebalancing may have been anticipated by market participants. Nonetheless, the sharpness of the decline suggests a more fundamental re-evaluation.

Despite the recent sell-off, some analysts maintain a cautiously optimistic outlook. Projections indicate that spending on AI infrastructure could exceed $400 billion by 2028. This long-term forecast suggests continued demand for high-performance chips. However, the market is now more acutely focused on the execution risks and the realistic timelines for revenue generation from these large-scale investments.

The current market sentiment reflects a shift from speculative enthusiasm to a more rigorous assessment of fundamentals. Investors are increasingly scrutinising balance sheets and cash flow statements, seeking concrete evidence of return on invested capital rather than relying solely on growth narratives. This trend suggests that future gains in chip stocks may be contingent upon demonstrated profitability from AI initiatives, rather than merely the promise of technological advancement. Whether current AI infrastructure outlays will yield the expected long-term dividends remains a critical question for the industry.