NVDA at 25x Earnings Bank of America Calls It a Rare Entry Point

 TechStockTracker
 
  TechStockTracker
 
Follow  Message
 
 
 
TechStockTracker decode the complex world of finance and investments, with a special emphasis on the dynamic intersection of technology and dividend growth stocks.
 

•  01 Dec 2025
 

•   AVGO  
   
  -1.19%
•   NVDA  
   
  +0.14%

There is a moment every serious investor watch for when a dominant company trades at a valuation that just doesn't match its position in the market. Bank of America thinks that moment has arrived for $NVDA and their analysis presents a compelling case that deserves attention.

Nvidia currently trades around $177 with a forward earnings multiple of approximately 25. According to Vivek Arya and the Bank of America analyst team, this valuation level has historically marked attractive entry points. Their research shows that each time Nvidia approached a 25x multiple, the stock recovered to levels between 30 and 40 within three to six months. Their price target sits at $275, implying roughly 55% upside from current levels.

What makes this particularly interesting is the broader context. Last week, the SOX semiconductor index surged about 10%, breaking back above its 50-day moving average and significantly outperforming the S&P 500 3.7% gain. Yet Nvidia itself dipped about 1% during this period, suggesting the rally broadened to other chip names while the leader lagged. This divergence often creates opportunity.

The investment narrative around artificial intelligence has shifted in recent months. Earlier concerns centered on whether cloud giants would maintain their massive capital expenditure levels. Now the conversation has moved to competition OpenAI versus Google Gemini in language models, and Google TPUs paired with Broadcom versus Nvidia GPUs in chip architecture. This shift in focus has weighed on Nvidia valuation despite its entrenched position.

Bank of America addresses the competitive concerns directly. They point out that chip supply remains tight across the industry, and Nvidia controls allocation. This makes it difficult for customers to meaningfully shift their purchasing mix over the next year. Additionally, while Google TPUs have proven themselves within Google own data centers, GPUs remain available across multiple cloud environments including Google Cloud itself. The bank estimates Google GPU purchases will reach around $10 billion in 2025, roughly matching their TPU purchases from Broadcom.

The technical comparison between Nvidia upcoming Blackwell Ultra chips and Google Ironwood v7 TPUs reveals nuances that matter to actual deployment decisions. Nvidia GB300 features 208 billion transistors compared to over 50 billion in the TPU v7 and carries 288GB of HBM3e memory versus 192GB for the TPU a 50% advantage. At the system level, Nvidia can scale up to 576 GPUs per pod while TPU v7 can reach 9,216 chips per pod, but with different performance characteristics depending on workload.

The cost analysis shows TPU v7 with about 45% lower total cost of ownership for FP8 workloads, but Nvidia leads in FP4 workloads where it doesn't have native support. More importantly, as Bank of America notes, while TPUs may be optimized for a limited set of tasks, GPUs maintain leadership across most AI workloads due to their established position and broader ecosystem support. This ecosystem advantage the software, tools and developer familiarity often matters more than raw specs on paper.

Looking at relative valuation, Nvidia now trades at roughly a 40% discount to $AVGO Broadcom earnings multiple. This represents the widest gap ever between the two companies, compared to historical averages where Nvidia traded at a 10% discount or even a 7% premium. The market has effectively shifted at least 10 percentage points of implied AI market share toward Broadcom for late 2026 and 2027. Bank of America views this as overdone, reminiscent of the DeepSeek moment in January 2025 that created what they saw as an attractive buying opportunity.

The analog chip space also caught a tailwind last week, with names like Analog Devices jumping 14% on strong earnings and raised guidance. Power supply for AI and optical components have emerged as additional growth drivers for select analog stocks. Nvidia itself has highlighted 14 different partners in the power supply space, and Bank of America expects this to become a secular growth market. This broader strength in chip-related names suggests the AI infrastructure buildout remains on track.

What's the bear case? Risks include potential weakness in consumer gaming, intensifying competition from both public companies and internal projects at cloud providers, larger-than-expected impact from China export restrictions, volatility in new markets, and increased government scrutiny of Nvidia dominant position in AI chips. These concerns are real and shouldn't be dismissed. The competitive landscape is genuinely evolving, and Google combination with Broadcom represents a formidable challenge.

Yet the analyst team maintains their buy rating with conviction, viewing the current multiple as a temporary dislocation. Their $275 target assumes a 28x multiple on 2027 earnings net of cash, within Nvidia historical range of 25 to 56. For Broadcom, they maintain a buy rating with a $460 target versus the current $403 price based on a 33x multiple for 2027.

The setup here is straightforward but not without risk. You have a company that dominates its market, trading at what historically has been a floor valuation, facing genuine competitive concerns that may or may not materialize as severely as feared. The chip supply dynamics favor the incumbent, the ecosystem remains Nvidia strength, and the total addressable market continues expanding Bank of America sees Nvidia reaching $1.2 trillion by 2030 three times current levels.

View All Headlines