Nvidia earnings countdown! A beat is almost certain, but Wall Street is most concerned about these five questions.

NVIDIA earnings season, the most important thing is no longer the numbers themselves.

On May 18, Bank of America Securities analyst Vivek Arya’s team released a preview report for NVIDIA’s Q1 earnings, which will be announced after the close on Wednesday, May 20, Eastern Time.

Based on NVIDIA’s historical performance over the past ten quarters, actual revenue has averaged 7% to 8% higher than management’s guidance. The previous guidance from management for F1Q27 revenue was $78 billion. According to this calculation, actual revenue is likely to fall in the range of $83–$84 billion, while the current market consensus forecast is only $78.7 billion.

In other words, a “beat” is almost certain. But analysts believe that the real market-moving issues after the earnings release are the following five questions.

Cash Return: Can NVIDIA’s "stinginess" change?

This is the topic most discussed in the report, and is considered the core reason for NVIDIA’s long-term valuation discount.

NVIDIA is currently the largest company in the S&P 500 by market value, accounting for as much as 8.3% of the index, surpassing Apple (peak 7.9%) and Microsoft (peak 7.2%) at their respective highs. The problem, however, is that NVIDIA’s shareholder returns are seriously mismatched with its scale.

The data is straightforward: from 2022 to 2025, NVIDIA’s free cash flow return rate (dividends + buybacks) averages only 47%, while the industry average in the same period is 80%, and even NVIDIA’s own ten-year historical average is 80%.

At the same time, NVIDIA’s current dividend yield is only 0.02%, while peers average 0.89%. Among equity income funds, only 16% hold NVIDIA, compared with 57% holding Microsoft and 32% holding Apple.

Where did the money go? Analysts point out that NVIDIA has invested large amounts into its ecosystem—OpenAI, Anthropic, tech partners. These investments are seen as controversial by outsiders; some believe it’s “circular financing”, meaning NVIDIA lends money to its customers, who then use the funds to buy NVIDIA chips.

How big is the valuation discount? Data shows that NVIDIA’s expected P/E for 2026/2027 is 26x/19x, while the average for the other “Magnificent Seven” members is 49x/42x—a discount close to 50%.

On a more specific comparison: Analysts predict NVIDIA’s combined free cash flow in 2026+2027 will exceed $430 billion, higher than the combined ~$375 billion of Apple and Microsoft. But NVIDIA’s market cap is about $5.46 trillion, about 28% lower than the combined market cap of Apple and Microsoft ($7.5 trillion).

Analysts believe that if NVIDIA increases its dividend and buyback intensity, it may attract more yield-focused long-term capital, narrow the valuation discount, and also dispel doubts about “circular financing.” They list this change as a “potential catalyst for the second half of the year.”

Vera Rubin: When will the next generation chip arrive?

NVIDIA’s current main product is the Blackwell series. The market’s concern: When will the next-gen Vera Rubin platform be officially scaled up?

Their judgment is the second half of 2026. Vera Rubin (codename R200) adopts TSMC’s 3nm process and shares the “Oberon” rack architecture with Blackwell Ultra, making product transitions relatively smooth and with limited impact on gross margins.

Looking further ahead, Vera Rubin Ultra (codename VR300) will be launched in the second half of 2027, adopting a completely new “Kyber” rack architecture, with high-bandwidth memory (HBM) accounting for an even greater portion of costs.

The market also wants to hear NVIDIA’s latest stance on the “trillion-dollar revenue forecast” in the earnings call—NVIDIA previously projected cumulative revenue of $1 trillion for 2025–2027, but the contributions from LPU (Language Processing Unit) racks, CPUs, and Vera Rubin Ultra were not included. Will there be an update this time?

Gross Margin: Can the 75% defense line be held?

Gross margin is one of the core supports for NVIDIA’s valuation.

Analysts judge: In the short term, since Vera Rubin uses Blackwell’s rack architecture, product transition gross margin will be relatively stable. But in the medium and long term, the rising share of HBM memory costs is an ongoing source of pressure.

Consensus expects NVIDIA’s gross margin to fluctuate between 74% and 75%. The bank agrees with this, but stresses that any gross margin performance above expectations would be a positive catalyst.

How will the AI accelerator market size forecast be updated?

Bank of America previously gave a “trillion-dollar” forecast framework for NVIDIA’s AI market in 2025–2027. This earnings call, the market will focus on whether NVIDIA updates that forecast, especially to include three new growth drivers previously excluded:

1. LPU (Language Processing Unit) racks
2. Vera CPU (NVIDIA’s self-developed server CPU)
3. Vera Rubin Ultra

The bank forecasts that by 2030, the overall AI accelerator market size will reach about $1.17 trillion, with NVIDIA maintaining about 68% to 70% market share.

Specifically, NVIDIA’s AI accelerator revenue is expected to grow from $102.2 billion in 2024 to $800 billion in 2030. AMD’s revenue will go from $5 billion to $80.1 billion, and Broadcom from $9.3 billion to $181.9 billion in the same period.

Is the threat from Google’s TPU and CPU competition exaggerated?

Recently, there is a saying in the market: As AI enters the “Agentic AI” era, CPU will become more important than GPU, threatening NVIDIA’s moat.

The bank disagrees, giving two reasons:

First, NVIDIA’s self-developed “Vera CPU” will have new progress disclosed at the upcoming Computex conference, and its competitiveness in the independent CPU market should not be underestimated.

Second, in the already massively deployed Blackwell and TPU clusters, the CPU-to-GPU ratio is now 1:2, which does not correspond to the narrative that “Agentic AI will need more CPUs.”

Conclusion: Although the CPU market is large, there are many competitors (both x86 and ARM architectures have strong opponents), and NVIDIA’s dominance in the GPU/AI accelerator field is unlikely to be shaken in the short term. By 2030, NVIDIA is expected to maintain about 70% revenue share in an AI total addressable market above $1.7 trillion.

Valuation: The "tech king" at half price

Finally, back to valuation. The report points out NVIDIA’s valuation paradox directly with a group of data.

Using expected P/E for CY26/27, NVIDIA is at 26x/19x, while the “Mag-7” average is 49x/42x—NVIDIA is discounted by nearly 50%.

Using EV/FCF (Enterprise Value/Free Cash Flow), NVIDIA is at 28x/20x, while the Mag-7 average is 83x/59x—discounted by over 66%.

Using PEG (P/E to Earnings Growth Ratio), NVIDIA is at 0.41x, Mag-7 average is 2.61x, and S&P 500 overall is above 1.3x.

Bank of America maintains a “buy” rating and target price of $320, based on an expected P/E of 28x for CY27 (excluding cash), which is in the mid-low range of NVIDIA’s historical valuation band (25x–56x).

~~~~~~~~~~~~~~~~~~~~~~~~

The above wonderful content comes from Chasing Wind Trader Desk.

For more detailed interpretations, including real-time breakdowns and frontline research, please join [Chasing Wind Trader Desk ▪ Annual Membership]

Risk Warning and DisclaimerThe market involves risks and investment needs caution. This article does not constitute individual investment advice and does not take into account specific investment goals, financial situations, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investment based on this is at your own risk.