Broadcom shares plunge 15pc as AI chip forecast disappoints
While second-quarter earnings beat estimates, weak AI sales outlook for the near term triggers sharp sell-off in pre-market trading.

Broadcom shares fell 15 per cent in pre-market trading on Thursday after the semiconductor giant’s quarterly results failed to meet investor expectations for future growth. The decline follows a strong five-session rally that had seen the stock hit multiple all-time highs, with market participants having priced in robust performance ahead of the earnings release.
The chipmaker reported second-quarter revenue of $22.19 billion, slightly exceeding the consensus estimate of $22.13 billion, although some analysts had anticipated stronger figures. Earnings per share came in at $2.44, beating the estimate of $2.39. Artificial intelligence semiconductor revenue grew 143 per cent year-on-year during the period, underscoring the company’s deep integration into the current data centre boom.
Despite the strong historical performance, the market reaction was driven by a disappointing outlook for the near term. Broadcom projected third-quarter AI chip sales at $16 billion, missing the consensus estimate of $17.2 billion. The company also confirmed it would not raise its AI semiconductor sales forecast for 2026, signalling a potential plateau in growth momentum compared to the aggressive expansion seen in previous quarters.
For the current quarter, Broadcom expects total revenue of $29.4 billion, which remains above the consensus estimate of $28.61 billion. Chief Executive Hock Tan stated that the company expects AI momentum to continue into fiscal year 2027, reiterating guidance that AI semiconductor revenue will exceed $100 billion in that period.
The sharp sell-off highlights the sensitivity of the semiconductor sector to forward-looking guidance. Angelo Zino, senior vice president at CFRA Research, noted that the high expectations leading into the report contributed to the volatility. Broadcom serves major hyperscalers including Google, Meta, and AI developers Anthropic and OpenAI, benefiting from an industry where hyperscaler AI spending is estimated to reach $650 billion this year.


