BlockBeats News, June 25th. Micron Technology, Inc. released an explosive earnings report early this morning, dealing a heavy blow to the entire market's short thesis. The company's revenue for the third quarter of the 2026 fiscal year reached $41.456 billion, significantly higher than the market's expected $35.423 billion. The company expects fourth-quarter revenue to reach $50 billion, also well above the market's anticipation of $42.915 billion. Adjusted earnings per share were $25.11, higher than analysts' expectations of $20.86.

More importantly, the signal released in the conference call is crucial. Micron stated that the number of strategic customer agreements has increased from 1 in the previous quarter to 16, covering approximately 20% of DRAM shipments and about one-third of NAND shipments. Of these 14 agreements, calculated at the minimum contract price, the remaining term has a cumulative revenue of about $100 billion. CEO Sanjay Mehrotra stated that these agreements will "fundamentally transform" the company's business model.

This statement is the most critical part of this earnings report. In the past, Micron has long been seen by the market as a highly cyclical memory company, with profits following the sharp fluctuations in DRAM and NAND prices. However, with the continued growth of long-term agreements, Micron's revenue visibility, pricing power, and profit stability will all improve. The market now has reason to reposition it from a traditional cyclical stock to an AI infrastructure supplier.

The conference call also reinforced the narrative of "persistent scarcity." Management expects the industry's tight state to continue beyond 2027, even though supply is expected to gradually improve by 2028, a clear timeline for supply to meet demand has not yet been seen. The reason is that new semiconductor fab construction is large-scale, complex, and has a long lead time.

This set of information directly addresses the three issues Citigroup highlighted in its Micron flash review: supply and demand, long-term agreements, and gross margins. Citigroup maintained its buy rating on Micron after the earnings report, with a target price of $1200, believing that the company's sales, EPS, and gross margins all significantly exceeded expectations, and that HBM and AI server demand will continue to support price increases through 2027. Citigroup stated that the market will next focus on whether long-term supply agreements can enhance revenue and profit sustainability. The 16 strategic agreements disclosed in the conference call provide a positive answer to this question.

JPMorgan's report places the Micron event within a broader understanding of the AI memory cycle. The bank believes that although memory stocks have seen significant gains in the past few months and volatility may increase, the fundamentals still support a "higher for longer" uptrend cycle. The bank has previously emphasized that LTAs long-term agreements will be a key variable driving the revaluation of memory stock valuations. The number and amount of agreements disclosed by Micron serve as a strong confirmation of this assessment.

JPMorgan also raised a larger question: Memory is becoming the crux of the AI business model debate. According to its estimates, the share of AI memory in cloud provider capital expenditures could reach about 52% by 2026 and surpass 70% by 2027, far exceeding the previous levels of less than 20% for AI. This signifies the rising importance of memory but also implies that cloud providers must prove that AI revenue and efficiency gains are sufficient to support this expenditure. Micron's long-term agreement shows that customers are willing to secure the supply, but the market will continue to question: Can these customers' AI investments deliver the expected returns?

Citi, in its latest report, stated that AI data center expansion requires not only HBM, DRAM, and NAND but also faster networks and optical interconnects. Citi projects that the global optical interconnect market could reach $92 billion by 2028, with a compound annual growth rate of about 65% from 2025 to 2028; 800G, 1.6T, 3.2T optical modules, as well as CPO/NPO and silicon photonics, will be the key growth directions.

International Business Strategies (IBS) believes that the semiconductor industry will remain strong in the long term, but a temporary downturn driven by memory may occur in 2028 due to reasons such as memory capacity expansion, 4F2 technology migration increasing supply, and high prices suppressing demand for traditional uses like PCs, smartphones, and automobiles. IBS also mentioned that data center capital expenditures may slow down in 2028 as cloud providers will focus more on low-cost inference and capital efficiency, such as alternative paths like China's low-cost tokens, photonics technology, and Google TPUs.

Deutsche Bank does not entirely agree with IBS's pessimistic outlook on memory in 2028. Deutsche Bank still believes that the tight supply of advanced processes and AI accelerators will continue until the end of 2027, with TSMC remaining in a strong position. What is truly noteworthy is Deutsche Bank's market reminder: Expectations are already high for AI capex, semiconductor equipment, TSMC expansion, and memory prices, and more fulfillment is needed in the future.

Therefore, the market implications of Micron's earnings report this time are not just "good performance." It is more like a repricing trigger for the AI hardware chain: In the short term, memory shortages, price increases, long-term agreements, and strong guidance support Micron and the storage sector to continue the uptrend; in the medium term, attention is needed as AI infrastructure spending is shifting from GPUs to memory and optical communication; in the longer term, Deutsche Bank's 2028 risk warning to the market reminds that any hardware supercycle will eventually face the test of supply expansion, capital returns, and cost efficiency.

Overall, Micron's earnings report reinforces the view that the AI memory cycle is still on the rise, but the Wall Street discussion has shifted from "Is there demand" to "How long can this cycle last, can it be locked in by long-term agreements, and will 2028 be rewritten by supply and efficiency."