By Cornelius Zeeman & Ashin Daya,
Fairtree Portfolio Manager & Equity Analyst
Nvidia recently joined an illustrious group of listed companies when their market capitalisation surpassed one trillion US dollars in July 2023. The trillion-dollar club includes the likes of Apple, Microsoft, Saudi Arabian Oil, Alphabet and Amazon. Nvidia’s rise is remarkable, considering its market cap was only USD 330 bn one year ago.
Artificial Intelligence (AI) has taken the world by storm this year, with the launch of ChatGPT and other AI applications showcasing the immense potential for AI to drastically revolutionise the way we gather, process, and consume information. Nvidia has been at the forefront of this AI inflexion, as they have supplied the powerful semiconductor chips used in AI servers behind all AI technology.
Nvidia capitalises on first mover advantage and superior product offering.
In the past year, Nvidia’s share price has gained 219% compared to the NASDAQ 100 Index’s 29% gain (see Graph 1 below). The company has been rewarded for its stellar results, and there has also been much enthusiasm around the future growth profile of the company following the launch of Open AI’s ChatGPT in March 2023.
Nvidia’s first-quarter FY2024 results surprised the market by guiding for second-quarter revenues to reach USD 11 bn, which was well ahead of consensus expectations of USD 7 bn. Thereafter, in their second-quarter results, they delivered USD 13.5 bn of revenue, which was USD 2.5 bn higher than guided. In addition, they guided for third-quarter revenues to reach USD 16 bn, which is well ahead of the USD 12.4 bn consensus expectations. Nvidia’s share price outperformance has been on the back of two exceptionally robust quarters (see Graph 2 below).
Graph 1: Nvidia’s Outperformance vs Nasdaq 100 Index
Source: Bloomberg, Fairtree
Graph 2: Group Revenue
Source: Company filings, Fairtree
The heat is on.
Nvidia has capitalised on its first-mover advantage in Graphics Processing Units (GPUs) – dominating the industry with more than 90% market share by units shipped (see Graph 3 below).
AMD is Nvidia’s leading challenger in the GPU market and has been very competitive in GPU pricing. However, performance has lagged as they have been focusing on producing GPUs for gaming consoles and notebooks. Later this year, AMD could disrupt the market dynamics as it is set to launch its newest and most advanced MI300 GPU chip for datacentres, set to rival Nvidia’s most advanced H100 chip and could very well challenge the market leader for market share.
Another competitor, Intel, is a new entrant to the GPU market. Intel’s GPUs are not yet as competitive as Nvidia’s; however, Intel is investing heavily in GPU development.
Graph 3: GPU Market Share (Units Shipped) – Nvidia, AMD & Intel GPU
Source: Citi Research, Mercury Research
GPUs increasingly take share from CPUs in datacentres due to their parallel processing capabilities.
While CPUs are optimised for general-purpose computing tasks, GPUs excel at handling highly parallel workloads, common in data centre applications like machine learning, artificial intelligence, scientific simulations, and data analytics. In addition to being faster, GPUs are more energy-efficient than CPUs because GPUs are designed to perform specific tasks, while CPUs are designed to be general-purpose processors – this means that GPUs can perform their tasks with less power than CPUs.
Graph 4: Datacentre Sales Processor Share
Source: Jefferies Research
The threat of substitution by high-performing ASIC chips.
Another competitive force is the continual growth and development of Application Specific Integrated Circuit (ASIC) chips (see Graph 5 below). Although these chips can be expensive to design and manufacture, they can offer significant advantages over CPUs. For example, Apple’s proprietary A-Series and M-Series chips have become the hallmark of their iPhones, iPads, and Mac computers. These chips offer impressive performance leaps over off-the-shelf chips previously used in these devices.
Among Nvidia’s datacentre customers, significant capital has been invested to create custom proprietary chips. Google has produced the Tensor Processing Unit (TPU), a high-performance ASIC used for AI and machine learning workloads. Amazon has designed the AWS Graviton chip, widely used in its datacentres to handle large workloads. These companies continue to invest in improving their custom chips to decrease costs and gain a technological edge while satisfying their ever-increasing computing performance requirements.
Graph 5: Custom ASIC Revenue by End Market
Graph 6: Dedicated Accelerator Market Share
Source: Jefferies, Technology and Semiconductors Research, September 2023, pg. 91
Nvidia is earning Gross Profit Margins of 70%, which is significantly higher than competitors AMD and Intel (see Graph 7 below). Naturally, high profit levels attract increased competition. As a result, Nvidia’s early lead in AI-related GPUs may well be challenged over the medium term as competitors launch their own GPU chips.
Graph 7: GAAP Gross Profit Margins: Nvidia, AMD and Intel
Source: Company filings, Fairtree
Nvidia is a quality company with an impressive history and a product offering. They are the first movers in the GPU industry and are benefitting from the early adoption of AI-based workloads. However, there are clear signs of emerging competition from other GPU producers, as well as custom chip solutions. While AI is undoubtedly an exciting technology which will significantly improve global productivity, we are still in the early stages of adoption and are yet to see a clear path towards widespread consumer adoption. This is an important prerequisite for sustainable demand, which is necessary to justify Nvidia’s trailing twelve-month P/E valuation multiple of 100 times.
Fairtree Asset Management (Pty) Ltd is an authorised financial services provider (FSP 25917). Collective Investment Schemes in Securities (CIS) should be considered as medium to long-term investments. The value may go up as well as down and past performance is not necessarily a guide to future performance. CISs are traded at the ruling price and can engage in scrip lending and borrowing. A schedule of fees, charges and maximum commissions is available on request from the Manager. A CIS may be closed to new investors in order for it to be managed more efficiently in accordance with its mandate. Performance has been calculated using net NAV to NAV numbers with income reinvested. There is no guarantee in respect of capital or returns in a portfolio. Prescient Management Company (RF) (Pty) Ltd is registered and approved under the Collective Investment Schemes Control Act (No.45 of 2002). For any additional information such as fund prices.