Nvidia (NASDAQ: NVDA) has established itself as the leader in the market for artificial intelligence (AI) graphics cards, boasting a staggering 92% market share. This dominance has fueled the company’s tremendous growth, with its stock soaring 230% in the past year. However, Nvidia now faces a potential threat from tech giants who are developing their own custom AI chips to reduce reliance on the graphics-card specialist.
Companies like Microsoft, Amazon, Meta Platforms, and Alphabet are Nvidia’s customers and have spent billions of dollars on its graphics processing units (GPUs). But these companies have been actively working on developing custom AI chips to address their specific AI workloads. For example, Alphabet has deployed custom AI accelerators called tensor processing units (TPUs) in Google Cloud. Meta Platforms plans to release its own custom AI chip this year, while Microsoft has already developed custom AI chips for deployment in its Azure data centers. Amazon has also introduced its own AI chips for training AI models.
There are two primary reasons why these tech giants are venturing into developing chips internally. First, Nvidia has struggled to meet the immense demand for its AI GPUs, resulting in long waiting periods for customers. Second, Nvidia’s AI GPUs are quite expensive, with the H100 processor costing between $30,000 and $40,000. In contrast, investment banking firm Raymond James estimates that it costs Nvidia just over $3,300 to manufacture one H100 GPU. This significant pricing disparity has motivated companies to explore alternatives to Nvidia’s offerings.
To address this threat, Nvidia is reportedly looking to enter the custom AI chip market. This move is not surprising considering the potential revenue opportunity. Investment banking firm Needham estimates that the overall custom chip market was worth around $30 billion in 2023, with AI commanding a significant portion of this space. Morgan Stanley predicts that by 2027, ASICs (application-specific integrated circuits) could account for 30% of the $182 billion AI chip market, potentially offering a revenue opportunity of $55 billion.
Nvidia is not alone in recognizing this opportunity. Leading ASIC manufacturers Broadcom (NASDAQ: AVGO) and Marvell Technology (NASDAQ: MRVL) have experienced a surge in AI-related orders. Marvell is projected to generate $1 billion in revenue from selling custom AI chips this fiscal year, while Broadcom is expected to sell custom AI chips worth $8 billion to $9 billion in 2024.
A former Marvell executive is reportedly heading Nvidia’s custom chip division, indicating the company’s seriousness about entering this market. It has already engaged in discussions with tech giants like Amazon, Microsoft, Meta, OpenAI, and Google to produce custom chips for them.
If the reports about Nvidia’s entry into the custom AI chip market are true, it could provide investors with another compelling reason to invest in this fast-growing AI stock. Despite its remarkable growth, Nvidia is currently trading at an attractive 35 times forward earnings, which is a discount compared to its five-year average forward price-to-earnings ratio of 42.
However, it’s important to note that investment decisions should be based on comprehensive research and analysis. The Motley Fool Stock Advisor analyst team, for example, recently identified their top 10 stocks for investors to buy, and Nvidia was not included in their selection. Therefore, individuals should evaluate their investment strategy and consider various factors before making any investment decisions.
In conclusion, Nvidia’s dominant position in the AI graphics card market is being challenged by the development of custom AI chips by tech giants. To counter this threat and tap into a potentially lucrative revenue opportunity, Nvidia is reportedly planning to enter the custom AI chip market. This move, if true, could further boost the company’s growth and provide investors with a compelling reason to consider investing in this AI powerhouse.
