The AI industry is now at a major tipping point toward mass adoption, as computing technology has improved and costs have plummeted. It is therefore not surprising that the AI industry is valued at $387.45 billion in 2022 and is forecasted grow at a rapid CAGR of 20.1%, reaching a value of $1.39 trillion by 2029. Picking a winner in the AI industry is challenging, so when there’s a gold rush, Why not sell blades by investing in a company like Nvidia? which has developed the backbone infrastructure through its revolutionary chips. In this post, I’m going to break down Nvidia’s AI advantage, its financials and valuation, let’s dive deeper.
What is AI?
Basically, AI uses software models to create human-like intelligence capable of completing tasks. Recently, the OpenAI institute, which is backed by Y Combinator’s Elon Musk and Sam Altman, has released a number of free AI tools that have gone viral. These tools include DALL-E, an AI image generator, which allows you to type any word in the search box and generate a custom image. For example, I asked the platform to create random images for me of a monkey wearing a hat and playing basketball, which it did.
I also created a “Warren Buffett wearing an apron” image, which you can see below. This tool may seem like a joke, but it has the potential to put many graphic designers and artists out of work.
Another popular tool advertised is ChatGPT, which offers a human language interface that can be used to answer exam questions and even write software code. The reason this is interesting is Nvidia (NASDAQ:NVDA) has established a strong position as a leader in artificial intelligence for enterprises, so if the aforementioned free tools sound interesting, imagine what the best supercomputers in the world can achieve, I’ll discuss that in the next section.
Artificial intelligence and Nvidia
Nvidia is a semiconductor chip designer known for being a Leader on high performance graphics cards. If you have a high-powered PC, chances are it uses an NVIDIA graphics processing unit. [GPU].
Nvidia also has a thriving AI business focused on the data center. Oracle recently announced an extension camaraderie with Nvidia to help scale AI for enterprises. Oracle cloud infrastructure [OCI] aims to make NVIDIA’s enterprise AI platform available to large organizations. As you can see in the chart below, Oracle’s cloud infrastructure uses 512 Nvidia GPUs per “cluster”. This equates to “tens of thousands” of Nvidia GPUs for all of Oracle’s infrastructure, this includes Nvidia’s A100 chips and the new H100 GPUs.
Nvidia’s H100 GPU is a technology discovery as it offers 7x better performance than your A100 for specific fast-paced computing tasks. Are include Genome sequencing, a rapidly growing industry. In addition to the Fast Fourier Transform [FFT]a mathematical model with applications such as RADAR and signal processing.
from nvidia AI platform enables organizations to develop AI solutions for their businesses. This includes all aspects of the process, from data processing to AI model training and deployment at scale. There are many applications including “conversational AI” bots, business process automation, computer vision technology, and much more. One specific use case is Nvidia Clara, which uses artificial intelligence and high performance computing. [HPC] for drug discovery, medical imaging, and even genomics.
Nvidia also has recently (December 7) announced a partnership with Deutsche Bank to accelerate the use of AI in the financial services sector. The bank will use Nvidia’s Enterprise AI platform on its Google Cloud infrastructure for a variety of applications. This includes the use of real-time “risk assessment”, which is generally a nightly process as it is CPU intensive. In addition, the bank plans to use it for “price discovery” and backtesting of models. This means your traders can manage risk more effectively while lowering energy costs with accelerated computing.
Deutsche Bank also aims to take advantage of Nvidia’s artificial intelligence platform to create virtual 3D avatars. They are expected to be used for a variety of applications, from customer service to navigating internal HR questions. The next stage after this will be “metaverse style” experiences with your banking customers.
Nvidia has also expanded its AI-powered “Digital Twin” business, which can be used to create copies of manufacturing facilities. This can be used to optimize production and better visualize improvements. Nvidia is even creating a “digital twin” of the complete land, which can be used to track global warming and possible changes in emissions.
nvidia reported mixed financial results for the third quarter of fiscal 2023. Its revenue was $5.9 billion, which was actually down 17% year-over-year. On the upside, this metric still beat analyst estimates by $114.74 million.
The revenue decline was primarily due to a substantial 51% decline in Gaming segment revenue to $1.574 billion. This may seem dire at first glance, it should be noted that the gaming industry has historically been cyclical and therefore a pullback in demand was expected after a big boom in 2020. Also, many consumers bought Nvidia graphics cards. for Bitcoin mining in 2020, and with the fall of Bitcoin we are also experiencing a “crypto winter”.
The data center is the engine of growth
One silver lining for Nvidia is its data center business, which has continued to thrive thanks to the aforementioned tailwinds on its AI chips. Its data center segment reported $3.83 billion in revenue, which increased a rapid 31% year-over-year. One positive for Nvidia is that this segment contributes ~64% of total revenue and therefore the fact that it is growing strongly is positive. This segment also continues to grow thanks to the ongoing “Digital Transformation” of businesses as companies move their IT workloads to the cloud. The cloud industry was valued at $429.5 billion in 2021 and is forecast to grow at a rapid CAGR of 15.8% through 2028.
In the first quarter of 2023, Nvidia plans to supply its new H100 AI chips to all major cloud infrastructure providers, including AWS, Azure, and Google Cloud. In mid-November, the company reported a large partnership with Microsoft Azure, whose goal is to build a large AI supercomputer for its enterprise customers.
Nvidia’s automotive segment is also growing rapidly, having grown its revenue by an astonishing 86% year-over-year to $251 million. This is still a small section of Nvidia’s overall business, but it has great potential.
Many car manufacturers, such as Mercedes, Volvo and Audi, use Nvidia’s open AI computing platform for cars called “Nvidia DRIVE”. The goal of the platform is to help enable self-driving technology at scale, as most automakers have not developed this technology in-house (unlike Tesla).
Nvidia reported operating income of $601 million, which was down a substantial 72% year-over-year. This was mainly due to a 31% increase in operating expenses, which compounded the decline in revenue, driven by cyclical demand for games. The good news is that most of these expenses appear to have been driven by a series of data center infrastructure expenses and increased headcount. So the company is really investing in itself and the data center expenses really should be considered “capex” as they are forecast to be of long-term benefit.
The company also has a strong balance sheet with $13.1 billion in cash and marketable securities. The company has long-term debt of $9.7 billion, but only $1.2 billion is current debt, which is manageable.
To value Nvidia, I plugged the most recent financial data into my discounted cash flow model. I’ve forecast 10% revenue growth for the next year and 26% revenue growth over the next 2-5 years. I predict that gaming demand will pick up and the data center segment will continue to grow.
To increase the accuracy of the valuation, I have capitalized R&D expenses, which has increased the operating margin to 33.25%. I have also forecast a pre-tax operating target of ~45%. I forecast this will be driven by operating leverage in the data center segment and a rebound in gaming.
Taking these factors into account, I get a fair value of $179.40 per share, the stock is trading at $163 per share at the time of writing and is therefore ~95 undervalued.
As an additional data point, Nvidia is trading at price/sales = 14.49, which is ~14% cheaper than its 5-year average.
Many analysts are forecast a recession, which can cause a lag in consumer and business spending. Demand for games may remain suppressed, while indirect revenue from cryptomining apps may never recover. Ethereum has moved from a proof of work to a proof of stake model, which is expected to have an impact on this.
Nvidia is a technology powerhouse and a true leader in data center hardware. The company is poised to benefit from the growth of the cloud and artificial intelligence industry. It is currently facing a number of headwinds due to low demand for games, but given that stocks are inherently undervalued, it could be a great long-term investment.