Nvidia NVDA is scheduled to release its fourth quarter revenue report on February 26th. This is Morningstar’s view on Nvidia’s earnings and what to look for in stocks.
Nvidia’s major morning star metrics
Revenue Release Date
Wednesday, February 26th, after the end of trading
What to look at Nvidia’s fourth quarter revenues
All eyes remain in the business of the data center. NVIDIA is on a healthy streak that reports results ahead of quarterly guidance, prior to Factset Consensus estimates. That said, the “beat” of the last quarter was less impressive than the “beat” of the previous quarter. This has led to modest concerns about slowing down AI spending.
I’ve heard from major Nvidia customers that CAPEX was significantly increased in 2025 (probably 30%-35% growth), but that Nvidia will win about a third of such spending I continue to look forward to this. It is a modest concern in light of US regulations and restrictive geopolitical AI landscapes. Nvidia’s AI chip business. Our model assumes that Nvidia will sell everything that they can build for the next 6-12 months.
The long-term data center discussion persists: some observers have debated the lifespan of scaling methods (pretraining AI models based on throwing ever-growing GPU clusters in the issue). Hyperscaler Capex suggests that these clusters are still being built, and Stargate’s announcement may have taken this debate a break in 2025 (although it would have put all $500 billion into investment). I’m skeptical).
Deepseek also looks like a game changer at the front of a Chinese-trained AI model that has less access to cutting-edge GPUs and appears to perform AI inference very efficiently. This poses the threat of fewer AI GPUs, large data centers and nuclear power plants.
Again, Hyperscaler Capex suggests that good investments are still moderate for Nvidia. However, the Slimmer AI model has always been part of the equation, and in our view, the arrival of Deepseek does not change Nvidia’s paper.
There are still important questions about the future of AI.
What AI workload is training vs. reasoning? Nvidia controls training, but says 40% of GPUs are used for inference. What does Nvidia look at on developing large-scale language models in the cloud? As genetics, drug discovery, robotics, or autonomous driving?
All these items inform our estimates of future data center revenue growth, resulting in free cash flow generation and fair value estimates.
Nvidia’s fair value estimate
With its three-star rating, we believe Nvidia’s shares are valued significantly compared to their long-term fair value estimates of $130 per share. This means a stock value of approximately $3.2 trillion. Our fair value estimates will double the price/adjusted revenue for fiscal year 2025 (end January 2025 or effectively calendar 2024) of 44 times and 30 times/adjusted revenue That means.
Our fair value estimates, and Nvidia’s share price, drives data center (DC) and AI GPU outlooks for better or worse. NVIDIA’s DC business has already achieved exponential growth, increasing from $3 billion in 2020 to $15 billion in 2023, and has since grown to more than $47.5 billion in 2024. It’s going online. Based on the strong results of NVIDIA in 2025 so far, it models DC revenue for fiscal year 2025 by 141% to $114 billion. It grew to $240 billion in fiscal year 2028, resulting in a CAGR of 23% for the next three years. The main driver of this growth is the ongoing increase in capital expenditures in data centers for large companies and cloud computing customers. We’ve only modelled 5% growth in 2029, as Nvidia could face an inventory correction or suspension of AI demand at some point in the medium term. It then has an average annual DC growth rate of 10%, which we consider as a reasonable long-term growth rate as AI matures.
Learn more about Nvidia’s fair value estimates.
Economy Moat Rating
Thanks to the intangible assets in the graphics processing unit, we allocate a wide moat to NVIDIA so that we can switch costs to our own software, such as the CUDA platform for AI tools.
Nvidia was an early leader and designer of GPUs, originally developed to offload graphics processing tasks for PCs and gaming consoles. The company has emerged as a clear market share leader for discrete GPUs (over 80% share per mercury research). We believe this leadership comes from the intangible assets associated with GPU design, as well as the associated software, frameworks, and tools needed by developers to use these GPUs. In our view, recent introductions such as the use of AI tensorcores in raytracing technology and gaming applications are indications that Nvidia has not lost its GPU leadership. A quick scan of gaming and DC GPU pricing shows that the company’s average selling price can be twice as high as its closest competitor, the advanced microdevice AMD.
Read more about Nvidia’s Economic Moat.
Financial strength
Nvidia is outstanding financial health. As of October 2024, the company had $38.5 billion in cash and investments compared to $8.5 billion in short-term and long-term debt. Semiconductor companies tend to hold large cash balances that help them navigate the chip industry cycle. During recession, this will provide cushioning and flexibility to continue investing in research and development necessary to maintain competitiveness and technology position. Nvidia’s dividends are virtually insignificant compared to its financial health and positive outlook, with most of the company’s distribution to shareholders taking place in the form of share buybacks.
Read more about Nvidia’s financial strength.
Risk and uncertainty
NVIDIA is assigned a very high uncertainty rating for Morningstar. In our view, Nvidia’s rating is tied to its ability to grow within the data center and AI sector, for better or worse. NVIDIA is the industry leader in GPUs used in AI model training, carving out a significant portion of the demand for chips used in AI inference workloads (running models to create forecasts or outputs) Includes).
We’ve seen many technology leaders compete for Nvidia’s leading AI position. It is inevitable that major hyperscale vendors such as Amazon’s AWS, Microsoft, Google and Meta platforms will try to reduce their dependence on NVIDIA and diversify their semiconductor and software supplier bases, including the development of internal solutions. I think so. Google’s TPU and Amazon’s training and recommended chips are designed with AI workloads in mind, but Microsoft and Meta have announced their semiconductor design plans. Among the existing semifinals, AMD is rapidly expanding its GPU lineup to serve these cloud leaders. Intel currently has AI Accelerator products, but it is likely that they will continue to focus on this opportunity.
Learn more about Nvidia’s risks and uncertainties.
The NVDA Bulls say
NVIDIA’s GPUs offer industry-leading parallelism, which has historically been needed for PC gaming applications, but is expanding to Crypto Mining, AI, and possibly future applications. AI model training. This is a use case that will rise exponentially over the next few years. It will move not only to industry-leading GPUs, but also to networking, software and services.
The NVDA Bears say
While Nvidia is today’s leading AI chip vendor, other powerful chip makers and high-tech Titans focus on in-house chip development, CUDA is the leader in AI training software and tools today, while major cloud vendors are This will hope for greater competition. If they occur, you may move to space and alternative open source tools. NVIDIA’s gaming GPU business is often watching the boom or bust cycle based on PC. Demand, and more recently cryptocurrency mining.
This article was edited by Gautami Thombare.
