The world’s once “hottest trade” in computer chips and memory has abruptly soured, with shares in major manufacturers plummeting. Companies in this sector develop essential components for modern electronics. These include data storage and processing units. On Tuesday night, this dive sent the Nasdaq 100 index, a key tech benchmark, briefly into correction territory, down over 10 per cent from its early June record. US memory giant Sandisk plunged 14.2 per cent, compounding Monday’s 11 per cent drop and halving its stock value this month. Other chipmakers experienced significant falls, contributing to the Philadelphia Semiconductor Index’s 25 per cent loss. Korea’s KOSPI index also dropped over 10 per cent, now 34 per cent down from its late June record.
This sharp sell-off stems from warranted profit-taking, notably after substantial gains like Sandisk’s 300 per cent rise since early 2026. However, the slump’s speed and violence are amplified by retail investors utilising margin loans and leveraged exchange-traded funds. Beyond profit-taking, the downturn highlights two key concerns for the artificial intelligence boom. Firstly, the risk that hyperscalers such as Microsoft, Amazon, Alphabet, and Oracle might slow data centre spending due to nervous investors. Secondly, worries that circular financing deals, such as Nvidia reportedly guaranteeing OpenAI’s massive data centre project, may be overextended.
Despite the tech sector’s turbulence, the broader market shows resilience. The S&P 500 index closed up 0.2 per cent on Tuesday and is just 2.3 per cent from its record high, up over 8 per cent for the year. This stability points to a significant market rotation, with investors moving away from expensive, overbought tech stocks into sectors offering robust earnings growth, strong balance sheets, and attractive valuations, notably healthcare and US financials. While this market broadening is positive, Bank of America analyst Nitin Saksena cautions such rotations do not preclude a longer-term AI bubble build. The speed and severity of the chip sell-off underscore potential broader market pain if hyperscaler AI spending truly retracts.
