Bank of America Analysts Say Chip Selloff Could End as Early as Q1 2027

Bank of America analysts, including Vivek Arya, say the semiconductor stock selloff could end as early as the first quarter of 2027, after a sharp drop that has already exceeded the average loss in four previous downturns.

The Philadelphia Semiconductor Index, known as the SOX, fell to 9,011 by Nov. 5 from a June 22 peak of 14,634, a 38% decline over 136 days. Four previous downturns, in 2011, 2015, 2018 and 2022, lasted an average 253 days and produced average peak-to-trough losses of 32%. The current decline has been steeper but has lasted just over half as long as the historical average.

In prior cycles, SOX declines generally came six to nine months before semiconductor sales weakened. The June 2026 peak could therefore precede earnings-per-share cuts in the first quarter of 2027, with policy changes affecting capital spending later in the year. Earlier downturns were followed by average rebounds of about 42% after earnings expectations had been absorbed.

U.S. cloud computing capital spending is running at a quarterly pace of about $300 billion, or $1.2 trillion annualized, compared with a consensus forecast of $1.555 trillion for 2027. If the current pace holds, spending next year would be about 23% below that forecast. For traditional hyperscalers, the current run rate is $982.5 billion against expected 2027 spending of $1.221 trillion, a gap of about 20%; emerging cloud providers have a gap of about 35%.

In a Bank of America scenario where 2027 capital spending comes in 20% below expectations, AI semiconductor revenue could face more than 20% downside pressure, while earnings per share could be at risk of falling 35% to 40%. The analysts said cuts could fall more heavily on land, power and facilities, while buyers continue to prioritize more efficient chips, including Nvidia's Vera Rubin.

Valuations have also compressed. The SOX trades at 13.7 times forward earnings, compared with 17.7 times for the S&P 500, a 22% discount. At the height of the AI rally, the index carried a 15% premium to the broader market; historical cycle lows have been near 12 times forward earnings. The analysts put the gap between the current multiple and those past lows at about 12%, while emphasizing that actual earnings reductions, rather than a low valuation alone, are needed to reassure investors.