AI chips surge as Bank of America forecasts $1.4 trillion market by 2030
Analyst Vivek Arya projects 100% growth in AI systems for 2026, boosts Nvidia and Broadcom estimates
Bank of America Securities has dramatically raised its forecast for the artificial intelligence semiconductor market, projecting the total addressable market for AI data-center systems will reach $1.4 trillion by 2030, up from a previous estimate of $1.2 trillion.
The revised outlook, published by analyst Vivek Arya, comes as cloud computing companies accelerate their capital expenditure plans and supply chain constraints ease. Arya now expects AI systems to grow 100 percent year-over-year in 2026, with overall data center systems growth projected at 64 percent annually as new accelerator deployments ramp up.
Nvidia stands as the primary beneficiary of the forecast upgrade. Bank of America raised the chipmaker's revenue estimates by 7 percent for fiscal year 2027, with earnings per share projections increased by 8 percent for that year and 3 percent for fiscal years 2028 and 2029. Arya described Nvidia as the "absolute leader" and "brain" of AI computing, citing strong market position and substantial sales visibility.
Broadcom received similarly bullish treatment, with revenue estimates increased by 5 percent for fiscal year 2026 and 2 percent for 2027 and 2028. Earnings per share were raised 4 percent for 2026, followed by 2 percent and 1 percent increases in subsequent years. Arya characterized Broadcom as the "nervous system" of AI infrastructure, highlighting its leadership in networking and optical interconnects alongside its custom silicon strategy for hyperscale cloud providers.
The forecast reflects broader industry confidence in AI's transformative potential. The global semiconductor market is projected to grow 15 percent in 2025, driven primarily by AI and high-performance computing demand. Meanwhile, the AI chip market specifically is forecast to reach $200 billion by 2027, up from $44 billion in 2023.
Recent data center performance supports the optimistic outlook. Nvidia's data center segment achieved record revenue of $51.2 billion in the third quarter of fiscal 2026, marking a 25 percent quarter-over-quarter and 66 percent year-over-year increase. Broadcom's AI semiconductor revenue surged 74 percent year-over-year to $6.5 billion in its fiscal fourth quarter, with management forecasting growth to $8.2 billion in the first quarter of fiscal 2026.
The companies' backlog figures underscore the sustained demand. Nvidia chief executive Jensen Huang has reported securing over $500 billion in orders for current-generation Blackwell processors and upcoming Rubin GPUs. Broadcom maintains a disclosed AI backlog of $73 billion scheduled over the next 18 months, including $10 billion in orders for AI switches alone.
Analyst sentiment remains overwhelmingly positive for both companies. Nvidia maintains a consensus "Buy" rating with an average price target of $264.20, suggesting 45 percent upside from current levels. Broadcom similarly carries strong "Buy" consensus with an average 12-month target of $433.87 from 34 analysts.
Arya downplayed near-term supply risks in his report, noting that a significant portion of expected 2026 high-bandwidth memory and advanced packaging supply growth is already allocated. The analyst also indicated that AI accelerator vendors can pass on rising costs to customers, preserving margins as sales volumes scale.
The AI infrastructure buildout is being funded primarily by strong cash flows from hyperscale cloud providers and government buyers rather than vendor financing, reducing financial risk exposure across the semiconductor supply chain.