Nvidia (NVDA) just became the first company in history to hit a $5.5 trillion market cap. Its CEO is flying to Beijing with President Trump. And on the same day all of that was happening, Bank of America’s top semiconductor analyst just changed a number that puts 45% upside on the table.
That number is built on a single assumption about the future of AI infrastructure that is considerably larger than what BofA believed just months ago.
What BofA changed on Nvidia stock and why
Bank of America analyst Vivek Arya raised his price target on Nvidia to $320 from $300 on May 13, maintaining a Buy rating. Nvidia shares traded near $220 at the time of the note, making the $320 target approximately 45% above the current price, according to Stocktwits.
More Nvidia:
- Nvidia is losing an industry that saved it from bankruptcy
- Nvidia CEO makes surprising admission on OpenAI and Anthropic
- Goldman Sachs just found a reason to like Nvidia stock again
The new target is anchored on an unchanged 28 times calendar year 2027 price-to-earnings multiple. What changed is the estimate base.
BofA raised its sales and EPS forecasts for fiscal years 2028 and 2029 by approximately 7% each, reflecting higher confidence in the multi-year AI demand trajectory, according to Investing.com.
The $1.7 trillion TAM number at the center of BofA’s thesis
The most consequential change in Arya’s note is not the price target itself. It is the market size assumption behind it.
BofA now projects the AI data center systems total addressable market will reach approximately $1.7 trillion by 2030, up from its prior estimate of $1.4 trillion, representing a 45% compound annual growth rate, according to Investing.com.
Within that, BofA raised its AI accelerator outlook to approximately $1.2 trillion from $1.0 trillion, driven by increased hyperscaler custom ASIC shipments including Google TPU and AWS Trainium. The data center server CPU forecast increased to approximately $110 billion, with AI CPUs accounting for $88 billion, up from $80 billion, Investing.com confirmed.
Arya believes 2026 “will continue to be a year of accelerating AI sales and ROIs,” while 2027 could see improving tokenomics and efficiency as new architecture compute and memory systems ramp, according to TipRanks. Companies positioned properly within the supply chain “should continue to deliver and outperform” despite ongoing bottlenecks, TipRanks confirmed.
Smialowski/Getty Images
What Nvidia’s Blackwell and Vera Rubin roadmap means for the thesis
BofA’s optimism is explicitly tied to Nvidia’s product roadmap.
The Blackwell architecture is already in volume production. The Vera Rubin platform, designed for the next phase of AI training and inference workloads, is expected to launch in the second half of 2026. Each transition represents a new demand cycle where customers need to upgrade infrastructure, which extends the revenue runway beyond what any single product generation can sustain.
The tokenomics argument is central to BofA’s 2027 view. As inference workloads scale, lower cost per token expands accessibility and compounds demand, a flywheel that Nvidia’s architecture roadmap is specifically designed to power. That dynamic, if it plays out as BofA expects, would mean AI compute demand grows alongside cost efficiency rather than being constrained by it.
BofA also flagged gradual gross margin pressure of approximately 30 to 50 basis points annually as high-bandwidth memory costs rise. That is a known headwind but a manageable one given the scale of revenue growth BofA is projecting.
Key figures from BofA’s Nvidia note and market context as of May 13:
- BofA new price target: $320, raised from $300, Buy maintained, analyst Vivek Arya, May 13; implied upside approximately 45%, according to TipRanks
- 2030 AI data center TAM: raised to approximately $1.7 trillion from $1.4 trillion; 45% CAGR, according to Investing.com
- FY2028 and FY2029 estimates: raised 7% each for sales and EPS; PT based on 28x CY2027 P/E, Investing.com confirmed
- Nvidia market cap on May 13: hit $5.5 trillion intraday, first company ever to reach that level, according to Stocktwits
- Nvidia YTD performance: up more than 21%; up more than 84% over the past 12 months, Stocktwits confirmed
- Q1 FY2027 earnings date: May 20; consensus revenue $78.98 billion, adjusted EPS $1.78, according to Investing.com
- Other BofA top picks in the note: Broadcom, Micron, AMD, and Marvell Technology with PT raised to $200 from $125, according to Investing.com
The context surrounding BofA’s note and what comes next
BofA’s raise arrives in the middle of a week when Nvidia is at the center of multiple converging catalysts.
Jensen Huang is accompanying President Trump to Beijing for the Trump-Xi summit, where a potential deal on chip exports to China could open a significant additional revenue stream for the company.
Separately, Nvidia’s Q1 FY2027 results are due on May 20, where consensus expects $78.98 billion in revenue and adjusted EPS of $1.78.
The Computex trade show in early June adds another near-term catalyst, with BofA noting the possibility of a new CPU launch that could broaden Nvidia’s addressable market beyond its core GPU business. Vera Rubin in the second half of 2026 completes the catalyst sequence for the year.
For investors, BofA’s note is a statement that the AI infrastructure buildout is not only continuing but expanding in scope beyond what the bank’s own prior models assumed.
The jump from a $1.4 trillion to $1.7 trillion TAM estimate is not a minor revision. It is a 21% increase in the addressable market BofA thinks Nvidia sits at the center of. How much of that market Nvidia ultimately captures is what the May 20 earnings report, and every report after it, will progressively reveal.
Related: Wells Fargo revamps Nvidia stock price target ahead of earnings