Nvidia’s Stock Is A Gift At $105 (Rating Upgrade)

Summary:

• Previously, I downgraded Nvidia Corporation stock to “Hold” due to stretched valuation, but after a 22% price drop and strong Q2 earnings, I upgrade it to “Buy” again.
• Despite high volatility and competition, Nvidia’s AI-driven growth, with a 122% YoY revenue growth, solidifies its position as the top AI player.
• Nvidia’s aggressive product cycle and strong CAPEX spending by major tech companies ensure continued dominance in AI GPU innovation despite minor setbacks.
• With a 12M forward P/E of 32x and a PEG ratio of 0.4, Nvidia’s stock appears undervalued, presenting a compelling buying opportunity.
• If Nvidia’s stock falls below $100, I will continue buying more shares.

I wrote an article on NVIDIA Corporation (NASDAQ:NVDA) just a few months back, downgrading the stock to a “Hold” as the price has been hovering near its all-time high of $140.

In my view, the price action has gotten ahead of the underlying fundamentals, particularly as growth slows due to the law of large numbers, not vice versa.

The downgrade has proven to be timely, with the stock down 22% since my previous coverage, with the broader market (SP500) remaining mostly unchanged.

Naturally, downgrading a hyped company like Nvidia is never a popular move among hardcore fans. Still, as investors, we need to be mindful of the price we pay in relation to the value we get.

Just to set the record straight, I have not sold any of my Nvidia shares. Instead, I added some more, around $90/share, during the broader market slump in August. In fact, Nvidia is among my top holdings right now, thanks to the strong year-to-date performance with the stock doubling.

I never try to time the market, particularly with momentum stocks like Nvidia, where the market may be more often irrational than rational; however, I am not willing to pay any price. Instead, with Nvidia’s strong growth, the maximum I am willing to pay is 25-35x its forward earnings.

Today, after Nvidia released its Q2 FY25 earnings in August, news surfaced about the upcoming delay of Nvidia’s Blackwell platform. The broader market has pulled back from the stretched valuation territory, the setup is once again favoring the bulls.

Let me walk you through the events of the past months and explain why I once again have upgraded Nvidia back to a “Buy” rating.

Top AI Play

Nvidia’s stock has been the number 1. beneficiary of the AI gold rush, with the stock up 108% on a year-to-date basis and 360% in the past three years alone, rewarding patient shareholders. At the end of the day, Nvidia’s stock has a 1.67 beta with an annualized volatility of 48%, either way.

To put it into perspective, Nvidia’s stock is 5.19x more volatile than the Dow Jones Industrial Average, perhaps not the best fit for every investor’s stomach.

Yet, not all semiconductors are created equal. Companies that did not jump on the AI bandwagon early enough or simply cater to different semiconductor uses, such as Intel Corporation (INTC) or Texas Instruments Incorporated (TXN), showcase very different stock performances.

With Nvidia’s business aimed to provide fundamental infrastructure through its Hopper and Blackwell GPU platforms to build state-of-the-art data centers enabling companies to train large language models (“LLM”) to unlock AI-enabled applications for their customers, I like to refer to Nvidia’s success as selling shovels during a gold rush.

Jensen Huang, Nvidia’s CEO, has recently become the new business superstar, taunting the success of the company’s business as humanity navigates a generational shift in computing:

The next industrial revolution has begun – companies and countries are partnering with NVIDIA to shift the trillion-dollar traditional data centers to accelerated computing and build a new type of data center – AI factories – to produce a new commodity: artificial intelligence.

Indeed, AI has become a new commodity. While it’s not a widely accepted fact among governments globally, private companies are racing to build the most advanced and immense AI data centers to stay at the forefront of innovation. They would rather overinvest now than underinvest and become obsolete in the fiercely contested tech field.

To name a few recent AI infrastructure investments:

• Amazon.com, Inc. (AMZN) is aiming to $11B on new data centers in Indiana
• Microsoft Corporation (MSFT) is investing $3.3B in Wisconsin
• Alphabet Inc. (GOOG), (GOOGL) is investing $1.1B to expand its data center footprint in Norway.

Now, there are some risks to Nvidia’s business model and AI in general as well:

• Most available public data has already been utilized to train large language models with diminishing returns and performance plateaus.
• Power requirements and water cooling consumption are enormous, prompting communities to oppose new AI data center development with an insufficient existing grid.
• Current transformer-based architectures may be approaching their limits, with fundamental changes in AI architecture that could allow true reasoning still years away.

Indeed, all these issues are real and potential threats to Nvidia’s superior growth. Nevertheless, the major tech companies are not yet showing signs of exhaustion, and their CAPEX spending continues to balloon to record levels.

For instance, Meta Platforms, Inc. (META) aims to have 350,000 Hopper (H100) GPUs in its computing infrastructure by the end of 2024 to continue developing its open-source AI model, Llama 3.1.

The estimated price for Nvidia’s H100 chips is between $25,000 and $30,000. If we assume Meta is paying the lower end of the range, with 350,000 chips, Meta alone plans to spend approximately $8.75B on Nvidia’s GPUs this year.

In 2024, AI represents the largest pie of CAPEX spend of many other Mag7 businesses.

That’s precisely the reason behind Nvidia’s explosive growth, with the revenue reaching $30B in Q2 FY25, growing 15% QoQ and 122% from the prior year.

As the competition does not come close to challenging Nvidia’s engineering superiority, particularly with Nvidia’s aggressive 1-year product cycle (now Advanced Micro Devices, Inc. (AMD) announced a 1-year product cycle as well), the company is enjoying monumental pricing power. This allows the company to capitalize on the opportunity while the rest of the industry catches up, with Gross Margin during Q2 FY25 at 75.1%, a slight contraction from 78.4% in Q1 FY25.

As a result of the aggressive product-cycle development, Nvidia’s operating expenses increased 48% YoY, but that’s easily offset by the combination of 122% top-line growth and 5.0ppts margin expansion from the period year.

Altogether, Nvidia brought in a record-breaking $16.6B in net income during the quarter, or a 1.68x increase from a year ago.

Following the earnings call, despite the record-beating results, which beat analyst estimates on both the top and bottom lines, Nvidia’s stock fell as much as 6.4%.

The reason behind the pullback is simply analysts’ lofty expectations. At the high end, they were hoping for as much as $37B in revenue, which has proved unrealistic, partially due to the law of large numbers, with growth deceleration on tap.

While historically, Nvidia’s business revolved around GPUs powering gaming, the narrative has shifted, and the business has reinvented itself. Data centers now make up 87.5% of the total revenue, while gaming and automobile segments have become minor parts of the business.

If you are confused about the large variety of AI chips that Nvidia currently markets, let me show you the most important ones:

• H100 GPU: Currently Nvidia’s flagship AI accelerator chip based on the Hopper Architecture
• H200 GPU: A variant of H100 explicitly designed for the Chinese market to adhere to US export restrictions.
• A100 GPU: Predecessor chip to the Hopper, built on Ampere architecture
• B200 GPU: The next-generation chip (not yet available), so-called Blackwell architecture.

Despite the high demand, the lead times of Nvidia’s core products reduced significantly from the previous 8-11 months waiting time to 8-12 weeks. The reduction in lead times can be partially explained by Taiwan Semiconductor Manufacturing Company Limited (TSM) allocating more production capacity to Nvidia.

Overall, Nvidia’s business is currently enjoying a reasonable degree of stability, with the Mag7 business making up 46% of its total Q2 FY25 revenue. Nevertheless, the concentration is real.

Let’s not forget that if the AI capex spending shows signs of weakness with one of the key players, that can be interpreted as a general pullback, with other players potentially reducing their spending. This is particularly true if the monetization question of AI will grow in importance from the investor’s side.

On the flip side, there are some signs that the AI tailwinds are widening. Governments, regional cloud providers, and smaller enterprises are all looking to invest in their own custom-specific AI infrastructure to elevate dependence on the top few players. There are increasing numbers of chatbots, copilots, and industrial solutions which could further boost Nvidia’s growth.

Likewise, while Nvidia currently holds 80% of the AI chip market share, the company is not the only player in town. AMD is playing catch-up, and cloud providers such as Amazon and Google all release their AI chips.

For instance, AMD’s product pipeline is similar to Nvidia’s 1-year product-cycle releases as a direct challenge:

• 2024: MI300 & MI325X.
• 2025: MI350.
• 2026: MI400.

So far, we cannot say with certainty that any of AMD’s AI chips have proven to be direct contenders, particularly as Nvidia’s business still enjoys industry-leading pricing power. However, the company has announced a delay in its Blackwell architecture due to a mask issue impacting the chip’s yield. The product ramp-up is now scheduled for Q4 FY25. This points to a three-month delay. It could cost Nvidia a few percentage points of market share and create an “air pocket” if customers such as Microsoft, Meta, Google, and Amazon wait on the new infrastructure instead of buying the current, Hopper one.

Heading into Q3 FY25, Nvidia expects to deliver $32.5B in revenue, representing another 79% growth. Gross Margins are expected to keep slightly contracting towards 74.4% to 75%.

Valuation

Given the impressive returns in 2023 and year-to-date in 2024, most investors would automatically avoid investing in Nvidia’s stock, claiming it must be too expensive.

Well, that’s partially correct, but in my opinion, it does not capture the whole story.

Nvidia’s trailing twelve months P/E is at a high of 46.5x.

However, it would be a shame not to capture inside the valuation the forward expected growth, which has so far proven to be spot on.

If we instead consider the expected growth of Nvidia’s bottom line, the narrative changes:

• FY25: Expected EPS of $2.81, 117% YoY growth.
• FY26: Expected EPS of $3.89, 38% YoY growth.
• FY27: Expected EPS of $4.70, 21% YoY growth.

From the growth expectations, we can see that Nvidia’s growth is indeed decelerating as a result of the law of large numbers, where it’s virtually impossible to keep growing 100%+ a year for a prolonged period with a revenue run rate of $100B.

Nevertheless, even if the decelerating growth materializes, Nvidia’s stock is priced at:

• Forward P/E FY25: 36.7x its earnings.
• Forward P/E FY26: 27.3x.
• Forward P/E FY27: 22x.

That’s a 12-month forward P/E of 32x its projected earnings, given that we still have two quarters to go in FY25 and two quarters to go in FY26.

Overall, 32x forward earnings is by no means a cheap valuation, but it’s a significant contraction from the trailing 12-month P/E and a valuation I am willing to buy Nvidia’s stock at.

Naturally, the future is inherently uncertain, and a slowdown in AI-CAPEX spending or a recession could hinder Nvidia’s strong momentum and growth. Nevertheless, at a 12-month forward PEG ratio of just 0.4, Nvidia’s stock appears to be undervalued.

If the stock contracts below $100/share or lower, I will scope the shares with both hands.

If the growth materializes and Nvidia’s stock trades at a P/E of 46.7x (15Y average) by the end of 2027, investors could see up to 23% annual ROR.

Takeaway

Nvidia’s reinventing of computing and the last few years’ stock performance are success stories I have never seen before and might never see again in my life.

The company has delivered a revolutionary product line-up at the right time to the market, which has the opportunity to unlock a new era of productivity and enable organizations and individuals to become more efficient.

Despite Blackwell’s platform’s three-month delay, Nvidia’s aggressive one-year product cycle positions it to remain at the forefront of AI GPU innovation with industry-leading pricing power.

The Mag7 business appears to continuously pour billions of dollars into CAPEX to build state-of-the-art AI data centers, ensuring their companies do not become obsolete, rather overinvesting than underinvesting.

Following Nvidia’s stock price slump and the tech-driven sell-off, I once again view Nvidia as a good deal, upgrading my rating from Hold to Buy.

Editor’s Note: This article discusses one or more securities that do not trade on a major U.S. exchange. Please be aware of the risks associated with these stocks.

---

Analyst’s Disclosure: I/we have a beneficial long position in the shares of NVDA either through stock ownership, options, or other derivatives. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Seeking Alpha’s Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.

---