Goldman Sachs Sees Buying Opportunity as Semiconductor Stocks Enter Oversold Territory

A recent report by Goldman Sachs indicates that the ongoing decline in semiconductor stocks may be excessive, presenting an attractive buying opportunity for investors. Despite near-term uncertainties regarding artificial intelligence infrastructure spending, the firm expects the sector to recover.

According to the report, the semiconductor index, or SOX, has experienced its sharpest relative decline against the S&P 500 over the past 10 years. The current drawdown is approaching 30 points, which is significantly higher than previous instances where the relative decline crossed 20 points. Historically, periods of such sharp underperformance are typically followed by phases of outperformance.

Goldman Sachs noted that stocks are clearly in oversold territory. The firm adjusted its assessments for current market conditions and analyzed relative valuations against potential downside scenarios across large-cap AI compute, memory, and semi-cap stocks. Although near-term catalysts for upward movement before the end of the year may be limited, the correction presents a favorable long-term entry point at the index level, even as broader market de-risking continues.

The investment bank identified two primary uncertainties currently weighing on semiconductor stocks: the funding gap between capital expenditure and free cash flows among hyperscalers and AI model developers, and questions regarding the future availability of sufficient land, power, and data centre construction capacity.

Despite these headwinds, Goldman Sachs projects robust growth in AI infrastructure spending. The firm estimates that total investments could reach approximately USD 1.3 trillion in 2027, marking a 60 percent year-on-year increase, before climbing to USD 2 trillion in 2028, up 42 percent.

Furthermore, AI-dedicated data centre deployments are projected to expand from roughly 20 GW in 2026 to 35 GW in 2027, and reach 57 GW by 2028. These investment estimates encompass spending by AI laboratories, sovereign AI enterprises, and corporate customers, extending beyond traditional capital expenditures by large technology companies.

Goldman Sachs characterizes the planned 35 GW of data centre additions for 2027 as relatively low-risk. For 2028, the firm expects at least another 35 GW of global additions to carry low execution risk, with potential deployments scaling up to 55-60 GW depending on project progression. However, physical constraints such as land and electricity availability remain critical factors that could influence expansion velocity and overall semiconductor demand.

Concluding its assessment, the investment bank reiterated that while immediate triggers for stock appreciation may be scarce prior to year-end, the overall risk-reward balance appears compelling for selective stocks at current valuation levels.

"Market corrections of this magnitude in foundational sectors like semiconductors often create strategic entry points for long-term investors. While near-term uncertainties around infrastructure funding and power capacity are valid concerns, the projected multi-trillion-dollar scale of AI infrastructure spending through 2028 points toward sustained structural demand. Investors must balance short-term market de-risking trends with long-term technological expansion." — Dr. Shishir Gupta, Founder & CEO, StartupLanes