Bank of America Securities: The AI industry is shifting from proving return on investment to deal with increasing chip and power restrictions in 2030 global semiconductor TAM to $3.2 trillion

The Zhitong Finance App learned that Bank of America Securities published a research report stating that it reiterated that the artificial intelligence industry is shifting from proving return on investment to dealing with structural and physical limitations such as chips and electricity; the shortage of memory chips and rising prices are still key levers driving the upward growth of the industry, and the prospects for artificial intelligence and servers are also more clear. The bank pointed out that the valuation of semiconductor stocks is still attractive, but it remains cautious before the midterm elections and until macroeconomic concerns ease.

The bank said that, driven by memory and data centers, and the recovery in the automotive and industrial sectors, it raised the global semiconductor industry's overall potential market (TAM) forecast for 2030 from US$2.7 trillion to US$3.2 trillion, or 18% CAGR between 2026 and 2030.

In terms of fab equipment (WFE), Bank of America Securities raised the 2026 WFE TAM forecast by 8% from the original US$144 billion to US$156 billion, up 33% year on year; the 2027 and 28 WFE forecasts were raised to US$210 billion and US$270 billion respectively, and reached US$360 billion by 2030, which means a compound annual growth rate of 23% from 2026 to 30.

The bank believes that although the market is concerned that investment in artificial intelligence infrastructure may slow down, customer orders, long-term agreements, capacity commitments, and semiconductor pricing have shown no signs of slowing down. The entire production capacity of computing power, network and memory suppliers in 2027 has been largely reserved or contracted, and it is expected that 2028 will remain tight.