Bank of America expects the AI data center market to grow fivefold to over USD 1.2 trillion by 2030. Against the backdrop of rapid market expansion, even if NVIDIA's market share declines from 85% to 75%, absolute gains will still experience explosive growth. Additionally, GPUs hold an irreplaceable advantage in public cloud and enterprise markets. Maintain a 'Buy' rating for NVIDIA, Broadcom, and AMD, as current valuations do not fully reflect their long-term profitability.
Despite the challenges posed by custom chips such as Google's TPU intensifying competition in the AI accelerator market, Bank of America (BofA) still recommends investors buy NVIDIA, Broadcom, and AMD amid the rapid expansion of the overall market.
On November 27, according to HardAI News, BofA stated in its latest research report that although Google is attempting to expand the ecosystem of its custom chips by leasing TPUs to Meta, which theoretically poses a challenge to NVIDIA and AMD, the bank maintains a strong bullish stance on NVIDIA, Broadcom, and AMD.
Analyst Vivek Arya emphasized that custom chips are mainly suitable for companies like Google and Meta, which have large-scale internal workloads, while GPUs remain irreplaceable in public cloud and enterprise markets.
BofA forecasts that the total addressable market (TAM) for AI data centers will grow fivefold by 2030, surpassing USD 1.2 trillion. Even if NVIDIA’s market share normalizes from the current 85% to 75%, its absolute gains will still experience explosive growth.
BofA recommends buying the three giants of the AI chip sector—NVIDIA, Broadcom, and AMD—across the board, believing that their current valuations do not fully reflect their long-term profitability.
Escalation in Model Wars: Are TPUs Beginning to Erode GPU Territory?
The AI hardware sector is currently undergoing a critical architectural competition.
The research report noted that Google has just released the Gemini 3 model, followed by Anthropic quietly launching Claude Opus 4.5 yesterday, claiming superiority over Google’s new model in coding and reasoning tasks.
BofA analysts pointed out that this indicates LLM development is a long-term marathon, and the current snapshot may not reflect the long-term evolution trend of market share.
Bank of America specifically highlighted that competition at the hardware level is even more critical. Google relies on its TPU, refined over a decade, for model training (both Gemini 2 and 3 are 100% trained using TPUs). Extremely crucial market rumors suggest that Google may lease TPUs to Meta as early as next year and potentially achieve local deployment cooperation with Meta by 2027.
If this information proves accurate, it will directly impact Meta’s current GPU suppliers—NVIDIA and AMD. This signals that custom chips (ASICs) are no longer merely proprietary tools for cloud providers but are beginning to demonstrate potential for external output.
Trillion-Dollar Market Explosion: NVIDIA's Normalized Market Share Does Not Alter Growth Logic
Despite the aggressive rise of TPUs, Bank of America argues that this does not alter the core logic of 'expanding the pie.' The bank provides three key reasons: explosive growth in market size, sustained share dynamics, and a robust moat, detailed as follows:
Bank of America forecasts that the total addressable market (TAM) for AI data centers will grow approximately fivefold from USD 242 billion in 2025 to over USD 1.2 trillion by 2030. Against this backdrop of rapid expansion, even if NVIDIA’s market share declines, its absolute revenue will still experience significant growth.
Bank of America's model assumes that as clients develop custom chips to handle diverse workloads, NVIDIA's market share will gradually normalize from the current 85% to 75%.
Bank of America believes that while custom chips offer cost advantages for specific internal workloads (such as Google and Meta's internal needs), they lack flexibility in public cloud environments.
Microsoft Azure, AWS, and over 100 emerging cloud service providers require high flexibility, making general-purpose GPUs an irreplaceable first choice. Even Google's own public cloud (GCP) continues to use NVIDIA’s GPUs.
Additionally, Bank of America analysts note that general-purpose GPU chips offer advantages such as spot availability, multi-cloud portability, a complete software stack, and a larger developer ecosystem. Moreover, tight supply chain conditions and NVIDIA’s economies of scale make it unlikely for market share to experience drastic changes in the short term.
Buy Across the Board: The Three Giants of AI Chips
The research report states that Bank of America has maintained its buy ratings for three chip stocks: NVIDIA, Broadcom, and AMD.
Despite NVIDIA's projected sales and earnings per share growth exceeding 40%, its trading multiple is only around 25 times. Bank of America considers this to be an undervaluation, maintaining a buy rating with a target price of $275.
Based on data and market share forecasts, NVIDIA’s earnings per share capacity is expected to exceed $10 by 2027 and surpass $20 by 2030.
Broadcom is seen as the biggest beneficiary of the custom chip trend. Bank of America assigns it a buy rating with a target price of $400, corresponding to a P/E ratio of 37 times for 2026.
Bank of America expects Broadcom’s AI business revenue in 2026 to achieve over 100% year-over-year growth due to additional TPU projects and the Anthropic initiative.
However, analysts caution that if Google directly licenses more TPUs, it could reduce Broadcom’s direct market opportunity for developing ASICs for other customers.
The report notes that AMD, as NVIDIA’s main competitor, remains a worthwhile holding. Bank of America assigns it a buy rating with a target price of $300, based on a 32-times P/E ratio for 2027.
This also reflects the company’s broad growth drivers across CPU, GPU, embedded systems, and gaming sectors.
Although there are risks associated with cyclical slowdowns in the embedded market, there remains significant upside potential in AI computing and gaining CPU market share.
As of Wednesday’s closing prices, the current stock prices of the three companies were $180.33, $397.57, and $214.24, respectively.
Editor/Wendey
