Bank of America Projects $2.2T Data-Center Market by 2030, Led by AI

Bank of America’s Global Research says the data-center market will reach $2.2 trillion by 2030, driven mainly by AI infrastructure spending. The forecast, led by semiconductor analyst Vivek Arya, is a continued upward revision from prior BofA estimates. Within the $2.2 trillion data-center market, AI data-center systems are expected to account for about $1.7 trillion, up from earlier ranges of $1.2T–$1.4T. BofA highlights a shift in chip demand: the server CPU total addressable market is now projected to exceed $210 billion by 2030, revised up from $170 billion. The report links this CPU rebound to “agentic AI”—systems that act with more autonomy to make decisions and execute tasks, not just generate text or images. These workloads are expected to rely more on CPUs than the GPUs used for training large language models. In BofA’s mix, the server CPU segment is about 10% of the overall data-center market, up from roughly 7% in earlier forecasts. Spending momentum is tied to hyperscalers (Microsoft, Google, Amazon, Meta). BofA expects their combined capex to exceed $700 billion in 2026, a 75% year-over-year increase. The bank also says it has financed over 5 gigawatts of data-center projects in the past 18 months, including a $16 billion Michigan campus for Oracle and OpenAI. Overall, the article frames a growing, AI-heavy supply chain across servers, accelerators, memory, networking, and related hardware—supporting a multi-year buildout of the data-center market.