Micron, the world's third-largest memory semiconductor corporation, is experiencing growing market expectations ahead of its fiscal year 2025 third-quarter (March to May) earnings announcement on the 25th (local time). Micron is the first in the memory semiconductor industry to announce quarterly earnings, acting as a gauge for industry trends. Analysts believe that thanks to the strong performance of high-bandwidth memory (HBM), a key component in artificial intelligence (AI) semiconductors, Micron is expected to significantly improve its earnings compared to the previous quarter.
According to financial information provider LSEG on the 23rd, global securities firms project Micron's third-quarter revenue to be $8.885 billion (about 12.28 trillion won). This slightly exceeds the midpoint of its guidance (self-projected estimate) of $8.8 billion. The adjusted earnings per share (EPS) estimate is $1.59, an increase of over 150% compared to the same period last year ($0.62).
As market expectations rise, major investment banks like Citigroup and Baird have raised Micron's target stock price to above $170. Micron's stock price, based on the closing price on the 20th, stands at $123.6, reflecting an increase of about 42% since the beginning of the year.
The bright market outlook is due to Micron's prominent position in the HBM business. As global AI server demand surges, the demand for HBM, which supports high-performance computing, is also increasing. Until last year, Micron's market share in HBM was only 5%, making it the clear 'bottom player.' However, this year, it has proven its technical capabilities in HBM3E (5th generation HBM), which has emerged as a key product in the market, shaking up the status quo.
In March, Micron began supplying its HBM3E 12-layer products to NVIDIA, which occupies 80% of the AI accelerator market. This positions Micron as the second player in the industry, ahead of Samsung Electronics. Samsung Electronics has not yet supplied HBM3E to NVIDIA. Micron's HBM sales have already surpassed $1 billion (about 1.38 trillion won) in the second quarter.
Preparations for next-generation products are also on track. On the 10th, Micron began supplying samples of its 36-gigabyte (GB) 12-layer HBM4 (6th generation HBM) to clients. HBM4 is expected to be fully integrated into AI semiconductors starting next year. Market research firm Trendforce forecasts Micron's HBM market share to reach 20% this year.
Sanjay Mehrotra, Micron's CEO, noted during the last earnings announcement that "thanks to AI-related demand, we have entered a trajectory to achieve record revenue and improved profitability in fiscal year 2025." To support the production capacity for next-generation HBM, Micron is intensifying investments in new fabs in Idaho and New York with the aid of government subsidies. Additionally, the company is focusing on strengthening its HBM competitiveness by appointing Mark Liu, the former chairman of TSMC, to its board.
However, concerns about declining profitability exist behind the rapid growth. In the second quarter of fiscal year 2025 (December to February), Micron exceeded market expectations for both revenue and EPS but reported a gross profit margin of 37.9%, falling short of expectations, which caused its stock to plummet nearly 8% in a single day. For the upcoming third quarter, the gross profit margin is expected to further decline to around 36.5%. Analysts suggest that intensified price competition and increased depreciation costs from new factory investments are straining profitability. JP Morgan stated, "We need to monitor whether this margin pressure is a temporary growth pain or a structural issue."
Therefore, the key point of interest in this earnings announcement will revolve around more than just the third-quarter performance figures; it will also depend on the fourth-quarter (June to August) guidance and the medium- to long-term strategies centered around AI memory presented by Micron's management. In particular, how specifically the plans for expanding HBM production capacity and securing profitability are presented will likely determine the market's confidence in the future growth of AI-driven memory.