Bank of America Names Nvidia (NVDA) Top Semiconductor Pick Ahead of Earnings
Key Takeaways
- •Bank of America reiterated a Buy rating on Nvidia with a $350 price target, citing the stock's most attractive forward price-to-earnings valuation in ten years at 16 times estimated 2027 earnings.
- •BofA projects fiscal second-quarter revenue of $94B to $95B, exceeding Nvidia's $91B guidance, and third-quarter revenue of $107B to $108B, surpassing consensus estimates near $104B.
- •The commencement of Nvidia's Vera Rubin chip deliveries marks the start of a multi-quarter upgrade cycle, while spot market GPU rental rates for B200 processors have risen to $5.66 per hour, indicating tight compute supply.
- •Although DRAM costs have climbed to 40–50% of total production costs from a historical 15–20% range, Bank of America expects only modest margin compression, forecasting gross margins to settle between 73% and 74%.
- •Nvidia has committed approximately $70B in venture investments across AI ecosystem partners, including $30B to OpenAI and up to $10B to Anthropic, reinforcing demand ties to leading AI model developers.

Shares of Nvidia (NVDA) opened Friday's session at $223.96, marking a 2.3% gain and trading near the stock's 12-month peak of $236.54. Despite this strong performance, Bank of America Global Research released analysis this week arguing that the semiconductor leader remains undervalued relative to its historical trading patterns. Nvidia's upcoming earnings report ranks among the most closely watched events of the quarterly season, as the company has become a bellwether for the broader artificial intelligence industry, with its results and outlook historically influencing sentiment across semiconductor, cloud computing, and AI software stocks.
According to the investment bank, Nvidia currently trades at just 16 times its estimated 2027 earnings per share — representing the company's most attractive forward price-to-earnings valuation in ten years. Bank of America reiterated its Buy recommendation alongside a $350 price target, designating Nvidia as its preferred selection in the semiconductor space. That distinction comes amid intensifying competition, with AMD advancing its MI300 accelerator series and major cloud providers including Google, Amazon, and Microsoft investing in custom silicon for AI workloads — efforts that have yet to materially erode Nvidia's dominant position.
Revenue Projected to Exceed Guidance in Q2 and Q3
Bank of America anticipates Nvidia will deliver fiscal second-quarter revenue in the $94B to $95B range, comfortably above the company's official $91B forecast. For the third quarter, the firm projects revenue between $107B and $108B, which would top the consensus estimate of approximately $104B. The demand backdrop remains robust, with major cloud and enterprise customers having publicly committed substantial AI infrastructure capital expenditures for the coming year.
This outlook stems partly from the commencement of Nvidia's Vera Rubin next-generation chip deliveries, which BofA characterizes as the start of an extended upgrade cycle spanning multiple quarters. Current GPU rental rates in the spot market have risen to $5.66 per hour for B200 processors, approaching all-time peaks — an indicator that available compute supply remains tight relative to demand. Bank of America projects Nvidia will maintain a commanding 65% to 70% share of the AI accelerator marketplace through the end of the decade.
Memory Cost Pressures Seen as Manageable
Among the primary concerns surrounding Nvidia ahead of its earnings announcement are escalating memory component costs. DRAM expenses now account for 40% to 50% of total production costs, a substantial increase from the historical range of 15% to 20%. The increase reflects growing reliance on high-bandwidth memory (HBM), which is essential for AI accelerator performance but produced by only a limited number of suppliers including SK Hynix, Samsung, and Micron.
Bank of America downplays these concerns, contending that Nvidia's strong pricing leverage and strategic supplier partnerships with companies such as SK Hynix will mitigate potential impacts. For Vera Rubin NVL computing racks specifically, memory cost increases translate to merely 60 basis points of margin pressure compared to the existing Blackwell Ultra platform. The bank expects gross profit margins to settle into a 73% to 74% band going forward, down modestly from approximately 75% currently. Key metrics investors will be watching in the earnings release include data center revenue growth, commentary on HBM supply allocation, and any updates to the Blackwell and Vera Rubin production timelines.
Venture Capital Activities and Financial Position
Bank of America also examined questions surrounding Nvidia's venture capital activities. The analysis indicates the company has committed approximately $70B in direct equity stakes across ecosystem collaborators, with $30B directed to OpenAI and potentially up to $10B to Anthropic. These investments deepen Nvidia's relationships with leading AI model developers while tying downstream demand for its accelerators to the fastest-growing AI workloads.
Given projected free cash flow of $469B spanning 2026 and 2027, BofA concludes that Nvidia's financial position remains robust.
When Nvidia last reported earnings on May 20, the company posted first-quarter revenue of $81.61B, representing 85.2% year-over-year growth, alongside EPS of $1.87 that exceeded the $1.76 consensus. Management also announced an $80B stock repurchase authorization and increased the quarterly dividend to $0.25 per share.
Wall Street consensus currently reflects a Buy rating with a mean price objective of $304.26. Bank of America's $350 forecast ranks among the Street's most bullish projections.