Stock Markets May 13, 2026 04:57 AM

Bank of America Lifts 2030 AI Data Center Market Estimate to $1.7 Trillion

Stronger forecasts for accelerators, networking and AI CPUs underpin higher valuations for key chipmakers

By Maya Rios NVDA MU AVGO AMD MRVL

Bank of America has revised upward its projection for the artificial intelligence data center systems market, now forecasting roughly $1.7 trillion by 2030 versus an earlier $1.4 trillion estimate. The firm raised forecasts across AI accelerators, networking and CPUs, and adjusted price targets for several semiconductor names based on a sum-of-parts valuation.

Bank of America Lifts 2030 AI Data Center Market Estimate to $1.7 Trillion
NVDA MU AVGO AMD MRVL

Key Points

  • Bank of America raised its AI data center systems market estimate to about $1.7 trillion by 2030, up from $1.4 trillion, implying a 45% CAGR.
  • Segment upgrades include AI accelerators to ~$1.2 trillion, data center server CPUs to ~$110 billion with AI CPUs at $88 billion, and AI networking to ~$316 billion.
  • The bank increased price targets and highlighted top picks including Nvidia, Broadcom, Micron, AMD, and Marvell, reflecting the larger market opportunity for AI hardware.

Bank of America has increased its outlook for the artificial intelligence data center systems market to about $1.7 trillion by 2030, up from a prior estimate of $1.4 trillion, implying a compound annual growth rate of 45% through the decade.

The bank highlighted expectations for accelerating AI sales and returns on that investment beginning in 2026. It identified upcoming initial public offerings from OpenAI and Anthropic as events that could improve market prospects. For 2027, the bank anticipates improvements in tokenomics and efficiency as new compute and memory architectures reach the market.

Specific segment-level revisions were significant. The AI accelerator market outlook was raised to roughly $1.2 trillion from $1.0 trillion, a move the bank attributes to stronger shipments of hyperscaler custom ASICs, including products such as Google TPU and AWS Trainium. Data center server CPU forecasts rose to about $110 billion, with AI CPUs now expected to represent $88 billion of that total, increased from a prior AI CPU estimate of $80 billion. The bank also lifted its AI networking forecast to around $316 billion from $240 billion.

Bank of America noted that as AI workload requirements broaden, diversification across compute and memory components will add to overall market value. The analysis envisions CPUs operating alongside existing GPU-CPU compute racks. It also expects SRAM-based ultra-low-latency memory racks to coexist with HBM-based GPU racks as different architectures serve distinct performance needs.

The research house updated company-level valuations based on these market changes. Micron Technology received a new price target of $950, up from $500, using a sum-of-parts approach that values the sustainable AI and HBM business at about $240 per share and the traditional DRAM and NAND business at roughly $710 per share.

The bank listed several top equity picks tied to the AI data center opportunity. Nvidia saw its price target raised to $320 from $300. Other top picks named include Broadcom, Micron, Advanced Micro Devices and Marvell Technology, with Marvell's target increased to $200 from $125.


Market context and implications

The revisions indicate a materially larger addressable market for hardware supporting AI workloads, driven by custom accelerators, increasing AI CPU penetration, and expanded networking requirements. The bank's stance links anticipated product rollouts and hyperscaler deployments to near- and medium-term revenue growth across semiconductor suppliers.

Risks

  • Projected acceleration in AI sales and returns is tied to expected IPOs from OpenAI and Anthropic; those events are cited as improving prospects but their timing and market effects are uncertain - this could impact technology and capital markets sectors.
  • Improvements in tokenomics and efficiency are anticipated in 2027 as new compute and memory architectures become available; delays or weaker-than-expected performance from these architectures could slow demand - affecting semiconductor and data center hardware markets.
  • Market expansion assumes diversification across compute and memory components; if customers consolidate around fewer architectures, some segments such as SRAM-based ultra-low-latency memory racks or certain accelerator ASICs may underperform - posing risks to suppliers in memory and accelerator markets.

More from Stock Markets

Porsche SE Families Demand Strategic Overhaul After First-Quarter Hit May 13, 2026 Investors Rotate Into AI-Linked Names as Momentum and Structural Growth Align May 13, 2026 Jakarta Stocks Slide, IDX Composite Hits New 52-Week Low as Market Loses Ground May 13, 2026 Adecco Shares Slide After Q1 Results Highlight Margin Shortfall Despite Revenue Growth May 13, 2026 Merck KGaA Rises After Q1 Beat and Upgraded 2026 EBITDA Guidance May 13, 2026