Key Moments
- Bank of America reaffirmed Nvidia as its top semiconductor pick, reiterating a Buy rating and a $275 price target following a “positive” investor relations meeting.
- BofA cited Nvidia’s GPUs as a full generation ahead of rivals, pointing to the transition from Hopper-based models to the upcoming Blackwell platform, which is described as “10x–15x better gen-over-gen.”
- Nvidia indicated visibility into “the $500bn sales outlook in CY25–26 at the minimum,” supported by strong utilization of existing GPUs and new partnerships with OpenAI and Anthropic.
BofA Reiterates Nvidia as Top Pick
Investing.com — Bank of America reaffirmed Nvidia as its preferred name in the semiconductor space in a research note Thursday, keeping its Buy rating and $275 price target intact. The decision followed what the firm described as a “positive” investor relations meeting with the company.
Technology Roadmap: Hopper, Blackwell, and Vera Rubin
The note identified Nvidia’s technological lead as a key driver of its investment thesis. Analyst Vivek Arya wrote that “NVDA GPUs remain a full generation ahead of competition,” emphasizing that the leading large language models trained on GPUs today are still built on the company’s 2022 Hopper architecture.
According to the report, Nvidia’s next major architecture, Blackwell, is on the way and is described as “10x–15x better gen-over-gen.” Blackwell-trained large language models are expected to appear in early 2026, extending what BofA views as Nvidia’s performance and capability gap versus competitors.
Nvidia also reportedly told investors that its subsequent platform, Vera Rubin, is progressing as planned. The company indicated that Vera Rubin remains on schedule for the second half of 2026, with a pre-fill inference CPX version anticipated in the fourth quarter of next year.
| Platform / Architecture | Key Detail | Timeline |
|---|---|---|
| Hopper | Current foundation for today’s leading GPU-trained large language models | Based on 2022 architecture |
| Blackwell | Described as “10x–15x better gen-over-gen”; expected to power next wave of LLMs | Blackwell-trained LLMs expected in early 2026 |
| Vera Rubin | Next-generation platform; includes a pre-fill inference CPX version | CPX in Q4 next year; full platform in 2H 2026 |
Customer Base and Ecosystem Strength
Arya underscored Nvidia’s customer positioning as a further pillar of the bullish stance. The note highlighted that “Google remains a key and growing customer, for NVDA every model builder still runs on NVDA.”
Beyond existing hyperscale relationships, Nvidia’s ecosystem was described as unique. The report pointed to what it called an “end-to-end platform… something no others can replicate,” arguing that this breadth supports the company’s competitive moat and monetization potential across hardware and software.
Demand Visibility, Utilization, and Partnerships
Bank of America wrote that Nvidia has clear visibility into robust demand. The firm stated that the company sees both demand and supply visibility into “the $500bn sales outlook in CY25–26 at the minimum.” The note added that newly formed partnerships with OpenAI and Anthropic represent incremental upside to that outlook.
Utilization metrics for older products also featured prominently. Nvidia told investors that even five-year-old Ampere GPUs remain “nearly 100% utilized,” supporting its internal view that a 5–6 year GPU life cycle is appropriate.
Margins and Valuation View
Despite pressures from higher memory costs, the report said Nvidia’s profitability expectations remain intact. The company’s “mid-70% GM outlook remains unchanged,” according to the note.
On valuation, Bank of America argued that Nvidia shares remain attractive. The firm characterized the stock’s valuation as “compelling,” indicating it corresponds to a PEG ratio of just 0.5x, which underpins its reaffirmed Buy rating and $275 price target.
TradingPedia.com is a financial media specialized in providing daily news and education covering Forex, equities and commodities. Our academies for traders cover 
