The demand for artificial intelligence (AI) chips went through the roof over the past year or so. Tech giants, start-ups, and governments are looking to get their hands on the most powerful chips available in the market to train AI models and deploy AI applications.
According to Deloitte, the market for generative AI chips could exceed $50 billion in revenue in 2024, up from almost nothing in 2022. Nvidia has been a big beneficiary of this booming demand for AI chips. Its revenue in the latest fiscal year 2024 is estimated to have increased to just over $59 billion, which would be a whopping 119% increase over the previous year.
What's more, Nvidia's dominant share of the AI chip market and the long-term growth opportunity illustrate why analysts are forecasting its earnings to double each year for the next five years. All this indicates that shares of Nvidia could keep heading higher following tremendous gains of 231% in the past year.
However, investors shouldn't miss the fact that there are other ways to capitalize on the AI chip market. While Nvidia is known for providing graphics processing units (GPUs) for training AI models, Arm Holdings (ARM +2.91%) and Micron Technology (MU +4.55%) also supply critical hardware and technological know-how to help accelerate AI adoption.
Let's look at the reasons why buying these two AI stocks instead of Nvidia could turn out to be a smart long-term move.
1. Arm Holdings
Arm Holdings stock experienced a major breakout of late following the release of the company's fiscal 2024 third-quarter earnings on Feb. 7.
Investors are buying Arm stock hand over fist since its latest quarterly report, and a big reason why is that Arm now joined the league of companies benefiting from the AI revolution. The British company, which doesn't manufacture any chips of its own but licenses chip architecture and designs to other chipmakers, saw a sharp jump in the number of Arm Total Access (ATA) agreements in the previous quarter thanks to AI.
NASDAQ: ARM
Key Data Points
Arm signed five ATA agreements in fiscal Q3. This took the company's total number of ATAs to 27. That's a nice jump considering that Arm reported 18 ATAs at the end of fiscal 2023 and had signed only four such agreements in the first six months of the current fiscal year. This improvement in the number of ATAs bodes well for Arm's future as it generates more revenue than other plans.
Jason Child, Arm's CFO, remarked on the latest earnings conference call, "As with recent quarters, we expect to sign multiple new ATA deals in Q4, and demand for our latest technology remains high as customers need access to AI-capable CPUs and related technology, such as our compute subsystems."
What's more, the adoption of AI led to a jump in demand for the company's latest Armv9 architecture, which it says commands double the royalty rates of the previous generation Armv8 architecture. That's not surprising as Arm designed the Armv9 architecture with an eye on improving AI processing capabilities such as voice recognition and image processing.
The Armv9 architecture accounted for 15% of Arm's royalty revenue in the previous quarter, which was an improvement over the architecture's 10% contribution in the previous quarter. As the demand for AI-enabled smartphones and personal computers (PCs) increases, Arm believes it could witness stronger demand for Armv9.
The good part is that the AI-driven interest in Arm's architecture and intellectual properties substantially improved its revenue pipeline. Its remaining performance obligations (RPO) stood at $2.43 billion at the end of the previous quarter, up 38% from the year-ago period. RPO refers to the amount of revenue that Arm will recognize in future periods, and this metric improved substantially last quarter as the company signed "multiple high-value, long-term deals."
This also explains why Arm's top-line growth is set to step on the gas following an estimated jump of 18% in the current fiscal year to $3.18 billion.
The company is also expected to deliver 41% annual earnings growth for the next five years. But don't be surprised to see Arm do better than that as more companies line up to sign long-term agreements and use its architecture to develop AI chips, which could help the company clock faster-than-expected growth in the coming years.
2. Micron Technology
The rapid deployment of AI servers created the need for more high-bandwidth memory (HBM). That's because HBM allows for faster transfer of data compared to traditional memory chips, which makes it essential for tackling AI workloads wherein large datasets are processed in real time.
This explains why top chipmakers such as Nvidia are looking to get their hands on as much HBM as possible to manufacture their AI graphics cards. In December last year, Taiwan-based publication DigiTimes reported that Nvidia paid $770 million to Micron as an advance to secure the supply of the latter's latest-generation HBM.
NASDAQ: MU
Key Data Points
Nvidia reportedly plans to deploy Micron's HBM3E product in its new AI chips this year. Comments from Micron's management on the company's December 2023 earnings conference call indicate something similar: "Micron is in the final stages of qualifying our industry-leading HBM3E to be used in NVIDIA's next generation Grace Hopper GH200 and H200 platforms."
Based on customer feedback, Micron points out that its version of HBM3E memory "has approximately 10% better performance and about 30% lower power consumption compared to competitive offerings of HBM3E." If that's indeed the case, it won't be surprising to see Micron cornering a nice share of the fast-growing HBM market.
According to market research provider Yole Group, HBM shipments could jump 147% this year, generating $14 billion in revenue. That would be a significant jump over last year's shipment growth of 93%. Even better, HBM shipments are expected to clock a compound annual growth rate of 45% through 2029, generating almost $38 billion in revenue at the end of the forecast period.
With Micron commanding an estimated 25% of the market for next-generation memory chips such as HBM, the company is in a nice position to take advantage of this emerging opportunity within the memory industry. Throw in the fact that Micron's top line is expected to jump 46% in the current fiscal year to $22.7 billion followed by another significant jump next year thanks to the improving memory market conditions, and investors would do well to buy it while it is still cheap.
Micron is trading at just 5.8 times sales, a nice discount to other AI stocks such as Nvidia that command an expensive price-to-sales ratio of 40. This makes buying Micron stock a no-brainer considering how fast it is set to grow and the lucrative AI chip opportunity it could take advantage of.
