Bank of America Securities recently released a report outlining the prospects for the semiconductor industry in 2026. Analyst Vivek Arya pointed out that the artificial intelligence boom continues to intensify and expand, and this decade-long industry transformation is currently only in its “middle phase.” He predicts that global semiconductor sales may see a year-on-year increase of 30% next year, thereby pushing the industry’s annual sales to surpass the historic milestone of one trillion dollars for the first time.
In the AI transformation, NVIDIA (NVDA) and Broadcom (AVGO) are seen as absolute leaders. Arya emphasized that NVIDIA’s development trajectory has entered a “league of its own.” Its graphics processing units (GPUs) are priced at around $30,000 each, which is entirely different from traditional chips that typically cost about $2.40 per unit. Although there are market concerns that its market capitalization may have peaked, Bank of America noted that NVIDIA’s free cash flow over the next three years is projected to reach $500 billion, and its price-to-earnings-to-growth (PEG) ratio is only about 0.6x, significantly lower than the S&P 500’s overall level of nearly 2x, indicating that its valuation remains attractive. If NVIDIA is considered the “brain” of the AI industry, Broadcom constitutes its “nervous system.” Broadcom has transformed into a core pillar of AI infrastructure, helping clients reduce dependence on single suppliers by designing custom application-specific integrated circuits (ASICs) for tech giants like Google and Meta. Goldman Sachs analysts believe Broadcom’s stock still has room for upside.
In addition to the two leading companies, Arya also listed Lam Research (LRCX), KLA Corporation (KLAC), Analog Devices (ADI), and Cadence Design Systems (CDNS) as top picks for 2026. He even pointed out that investors are unlikely to go wrong by simply ranking semiconductor companies by gross margin and buying the top five. These companies typically hold 70% to 75% market share in their respective segments, which is the norm for leaders in the technology sector.
Bank of America estimates that by 2030, the total addressable market (TAM) for AI data center systems will exceed $1.2 trillion, with a compound annual growth rate (CAGR) of 38%. Within this, AI accelerators alone represent a market opportunity of $900 billion. However, the path to a trillion dollars will be “bumpy.” The primary challenge lies in the extremely high construction costs of AI data centers. A standard 1-gigawatt AI data center requires capital expenditure of over $60 billion, with roughly half allocated to hardware procurement. This raises core questions about whether massive investments can deliver substantial returns. In response, Arya remains optimistic, believing that tech giants’ investments in AI represent a strategy that is both “offensive and defensive,” and that continued investment is a necessary choice to defend their business territories.
