The past few years have seen a notable increase in the shipment of artificial intelligence (AI) servers, as cloud service providers have been investing large sums of money in infrastructure that can train AI models and be used for AI inferencing to implement those models in practical applications.
The global AI server industry is expected to generate a staggering $187 billion in sales this year, growing 69% from 2023, according to market research firm TrendForce. Numerous businesses are already reaping significant benefits from this enormous end-market opportunity. There are several avenues to invest in the expanding AI server business, ranging from chip makers like Nvidia to makers of custom chips like Broadcom and server solution suppliers like Dell Technologies.
But in this piece, we’ll focus more intently on the futures of two businesses that produce essential parts for AI servers: Marvell Technology and Micron Technology.
Also Read: Flow Matching by Pyramid Flow: Create Smooth, High-Quality 10-Second Videos Efficiently
The high-bandwidth memory chips from Micron Technology are in great demand
AI server processors, such as graphics processing units (GPUs), employ high-bandwidth memory (HBM) because it facilitates quicker data transfers, which lower processing times, increase performance, and consume less power. According to Micron, it has sold out all of its capacity for this year and the next due to the high demand for HBM.
Better yet, Micron management notes that because of the new contract it has secured for its most recent HBM3E chip, it “will have a more diversified HBM revenue profile” for 2026. The chip manufacturer notes that it has started shipping this new chip to its clients for approval.
In comparison to competing products, Micron claims that HBM3E uses 20% less power and offers 50% more capacity. Early in 2025, the business plans to begin ramping up production of HBM3E, and as the year goes on, it will boost output. Better yet, Micron is certain that it will keep expanding its market share in the HBM sector.
In the current fiscal year (which began on August 30), Micron’s revenue is predicted to soar by a startling 52% to $38 billion, driven by the company’s focus on capturing a larger portion of the HBM sector and a growth of the end market. Analysts predict that Micron’s earnings will rise from $1.30 per share to $8.94 per share in the meanwhile.
It is anticipated that Micron will continue to expand at an impressive rate in the upcoming fiscal year.
Investing in Micron Technology at this time may prove to be a wise choice for anyone hoping to profit from the increasing use of AI servers. The stock is extremely cheap concerning the growth that is anticipated, as seen by its forward earnings multiple of just 11 and its price/earnings-to-growth ratio (PEG ratio) of just 0.16.
Also Read: Atlassian Rolls Out Rovo AI Assistant and Focus to Streamline Enterprise Planning at Team ’24
Marvell Technology’s bespoke AI processors are helping to propel the company forward
The company Marvell Technology is well-known for producing application-specific integrated circuits (ASICs), which are specialized chips made to carry out particular functions. It’s important to note that since big cloud service providers like Meta Platforms, Alphabet’s Google, and Amazon are trying to cut costs by creating in-house processors, there is an increasing demand for these custom chips used in AI servers.
Because of this, it is anticipated that in 2024, ASICs will represent 26% of the total market for AI server chips. Better yet, the use of ASICs in AI servers is anticipated to grow at a reasonable rate in the future, creating a $150 billion potential revenue opportunity. Marvell is already taking advantage of this significant change.
In the second quarter of fiscal 2025 (for the three months ended August 3), the company’s overall revenue decreased by 5% year over year to $1.27 billion, primarily due to difficulties in the carrier infrastructure, consumer, automotive, and business networking end markets. On the other hand, it produced an amazing 92% year-over-year rise in data centre income, reaching $881 million.
Marvell anticipates that as a result, the growth of its data centre business would “accelerate into the high teens sequentially on a percentage basis” in the current quarter. This is a step up from the 8% sequential growth the company announced in the prior quarter. Marvell’s projection for the upcoming quarter indicates an improvement in its financial performance, which makes sense.
Revenue in the company’s third fiscal quarter is expected to reach $1.45 billion, up from $1.42 billion in the same time the previous year. Marvell is therefore expected to resume growing from the current quarter onward, and analysts anticipate that it will increase strongly during the ensuing two fiscal years.
For the following five years, experts anticipate Marvell’s earnings to expand at a compound yearly growth rate of 21%. Thus, investors who would like to add Marvell Technology to their portfolios to capitalize on the increasing demand for custom AI chips should think about acquiring a semiconductor stock. Its expansion is expected to pick up speed because of the enormous potential in the AI server market.
Also Read: AMD’s New AI Chip Instinct MI325X Targets Nvidia’s Dominance in Data Centers
