A prominent American tech investor and founder of Nelson Investment Management, Dan Niles, has stated that the sharp decline in SK Hynix's stock price, along with the broader semiconductor sector, signals a "near-term floor" for artificial intelligence (AI) stocks. He attributes the intense sell-off to forced liquidations rather than a fundamental collapse of the industry.

Niles made these remarks after SK Hynix missed its second-quarter operating profit consensus and simultaneously lowered its fiscal 2026 capital expenditure guidance by 11%. While he acknowledges that the reduced capex forecast provides ammunition for investors who claim the AI cycle has peaked, Niles insists it is merely a short-term "speed bump." He notes that the recent market movements, driven by forced liquidations and margin calls on retail accounts and hedge funds, have already formed a "near-term floor" while accelerating market cleansing. This is because major prime brokers are keen to avoid a repeat of the Archegos Capital Management meltdown.

The recent semiconductor sell-off has been exceptionally brutal. Niles points out that since the peak on June 22, the S&P 500 index has only fallen 2.1% and the Nasdaq index has dropped 6.6%, but the Philadelphia Semiconductor Index (SOX) has plummeted 28.6%. Additionally, the Morgan Stanley Momentum Index has tumbled 38.0%, and the TMT Momentum Index has crashed 53.5%. Niles describes the magnitude of this decline as historically rare.

Nevertheless, Niles remains firmly bullish on the underlying fundamentals of the industry. He draws a parallel between the current market and the 1995 tech stock correction, when Windows 95 underperformed market expectations, forcing Intel to write down $1 billion in inventory. Today, he argues, "companies' balance sheets don't have a memory inventory glut issue." With the rise of agentic AI, a technology requiring immense computing power, Niles believes that while the top 1% of enterprises are finding ways to compress token consumption, the remaining 99% are still expanding their computing power usage. Therefore, he predicts that the sectors hit hardest during this brief correction will experience a powerful rebound.