In a recent analysis, JPMorgan has encouraged its clients to embrace the current selloff in semiconductor stocks, particularly singling out Broadcom as a 'Strong Buy' for the remainder of 2026. This advice comes amidst growing investor anxiety about sustainability in AI infrastructure spending beyond the current projection timeline.
The AI Chip Backlog as a Catalyst
JPMorgan's Mid-Year Outlook 2026 highlights a substantial backlog in AI chip orders deemed as pivotal to their investment recommendation. This dynamic suggests that the demand for AI chips is outstripping supply, providing a strong foundation for revenue visibility that could extend well into the future. The research team anticipates that hyperscale capital expenditure will rise by $130 billion, leading to a staggering total of over $650 billion in expected investments through 2026.
The Pressure on Semiconductor Stocks
Despite these bullish forecasts, semiconductor stocks have faced pressure recently, attributed to skepticism on Wall Street regarding the longevity of the hyper-scaling spending spree led by major tech firms like Microsoft, Google, and Amazon. Investors are debating what will transpire after this initial boom in infrastructure buildout whether spending will plateau or decrease significantly.
Risks and Opportunities
While JPMorgan has recommended a focus on semiconductor investments, they have not overlooked the inherent risks. The bank has noted a narrowing valuation gap between chipmakers and the hyperscalers that purchase their products, indicating a potentially shrinking margin of safety for investors entering the market now. This situation creates a compelling yet cautious narrative for future investments in this sector.
Moreover, the report reinforces that global equities could benefit from AI-driven productivity enhancements well into late 2026, with semiconductor companies positioned as key players in this evolving market landscape. Broadcom, specifically, emerges as a particularly intriguing prospect, given its dual involvement in custom silicon and networking infrastructures, which places it at a favorable vantage point in the AI chip race.
