A truly explosive financial report with a perfect score of 100 and a market expectation of 120, which actually scored 150 points. According to the forward-looking framework, Micron's financial report can be broken down as follows: 1. This quarter's revenue was $41.5 billion, far exceeding the market expectation of $38 billion, with a gross profit margin of 84.9. The four major business segments of data center, cloud storage, mobile client, and automotive embedded have all exceeded expectations, which means 1) The main theme of the current storage cycle is of course AI, but if only AI business is strong and traditional terminals are weak, the market will still worry about growth being too concentrated and price increases being difficult to spread. The unexpected performance of mobile and customer services indicates that the tight supply and demand of storage is spreading to more downstream areas. 2) At the same time, the product portfolio is continuing to tilt towards high-value direction. AI infrastructure requires storage solutions with higher capacity, higher bandwidth, and lower power consumption, which typically have higher ASP and stronger bargaining power. Simply put, Micron's financial report does not convey that "only AI is strong", but that "AI drives the improvement of the supply and demand structure of the entire storage industry". 2. Micron expects adjusted revenue of $49 billion to $51 billion for the next quarter, with a median of approximately $50 billion, significantly higher than analysts' expectations of $43.24 billion. According to the median calculation, the guidance is about 15.6% higher than market expectations. The most aggressive forecast for the next quarter mentioned in the previous financial report forecast was Goldman Sachs' 48.8 billion. It is equivalent to Micron's own guidance lower limit being higher than the most optimistic expectation. The company expects an adjusted EPS of $30 to $32 for the fourth quarter, with a median of $31, which is $25.31 higher than market expectations and about 22.5% higher than expected. The importance of this set of guidelines lies in its ability to dispel market concerns that current profits may be approaching their peak. Against the backdrop of a significant increase in Micron's stock price this year and extremely high market barriers, simply "meeting expectations" is not enough to drive the stock price to continue rising. What really drove the post market stock price up by more than 10% was the company's strong signal of further acceleration in next quarter profits 3. SCA/LTA Long Term Cooperative Mechanism Micron has signed 16 Strategic Customer Agreements (SCA) with customers in the data center, consumer end, and automotive markets, including 4 super large customers and 3 mid size customers, with key SCA terms. These long contracts are not simply intention agreements, but rather take or pay contracts with strong binding force. The agreement usually lasts for five years (from 2026 to the end of 2030), specifying the specific purchase quantity and setting the price bottom line and upper limit. These 16 long orders have brought highly impactful financial moat data to the market: 1) Guaranteed revenue of 100 billion US dollars: Based on the minimum price and minimum shipment volume agreed upon in the contract, the guaranteed revenue corresponding to future performance obligations (RPO) has reached 100 billion US dollars. 2) Huge cash flow advance: Customers have provided real money and silver to obtain stable production capacity supply. CFO Mark Murphy revealed that under the SCA signed so far, we expect to receive $22 billion in cash deposits and related financial commitments. The vast majority of it (approximately $18 billion) will be in the form of cash deposits. ” 3) Thoroughly smoothing cycle fluctuations: Currently, these long orders have locked in approximately 20% of Micron's DRAM capacity and 33% of its NAND capacity during this period. Mehrotra, Chairman of Micron, emphasized that for SCA that includes price ranges, even at the bottom price level, Micron's gross profit margin will be very generous, far exceeding the peak quarterly profit margin of any previous cycle. ” 4. HBM4/HBM4E Roadmap and NAND/eSSD Data Center Business 1) The high-intensity production speed of HBM4 12hi is twice that of HBM3E 12hi, and Micron has achieved over $1 billion in HBM4 revenue. The management of Micron expects that the maturity yield of HBM4 12 layers will be significantly faster than that of HBM3E 12 layers. 2) It is expected that future memory demand will continue to tilt towards higher performance and higher value products, whose complexity brings higher cost per chip. The transition from LP5 to LP6, DDR5 to DDR6, and the updated generation HBM are all accompanied by an increase in grain costs. This trend, coupled with the large-scale production of new capacity in the coming years, is expected to lead to an increase in the cost of DRAM hybrid chips from the current level It is expected that by 2026, the shipment volume of DRAM and NAND bits in industry data centers will more than double compared to two years ago. The expected increase in shipment volume is driven by a moderate decrease in the average DRAM content growth of servers, as customers focus on maximizing unit shipment volume amidst extremely tight memory allocation. In the NAND field, AI contextual memory storage and HDD (hard disk drive) replacement opportunities are expanding the addressable market of SSD (solid state drive). 5. Future opportunities and demand space 1) The market is highly concerned about the sustainability of this storage explosion. In response, Micron's management has provided a clear timetable judgment: scarcity is not only the current state, but also the norm for the next two to three years. Chairman Mehrotra stated that this dependency has led to a qualitative change in the role of memory: The performance of AI systems depends on the performance and capacity of the memory subsystem in terms of architecture. It has elevated the status of memory in the AI world, making it a strategic asset. ” 2) Why is it difficult to keep up with the supply? Micron identified four major structural pain points: Firstly, the construction of new wafer fabs is massive, time-consuming, and constrained by shortages in labor and energy infrastructure; Secondly, as nodes upgrade (such as One Gamma and G9 nodes), the increased complexity of the process leads to a slowdown in bit growth; Thirdly, HBM extremely consumes wafer production capacity, severely squeezing the supply of non HBM products; Fourthly, the limited clean room space restricts the expansion of production capacity 3) In addition to data centers, the imaginative space of edge AI and terminal devices is being opened up. In the automotive industry, L2+and above level autonomous vehicles have more than 5 times the memory and storage capacity of ordinary cars, and the penetration rate of such cars is rapidly soaring. The robot race track has a more long-term impact. Chairman Mehrotra has outlined a huge incremental blueprint for the market: The memory carried by humanoid robots is 10 times that of ordinary L2+level cars. We anticipate that in the latter half of this decade, a sustained, substantial, and decades long memory demand cycle will begin. ” Even in 2028, as supply begins to gradually improve, it is expected that demand will continue to maintain a strong trajectory, as these AI trends are very long-term trends. AI is still in a very, very early stage. The entire token economy (referring to the processing cost of AI models) requires more memory. The performance of AI systems is actually limited by memory capacity, memory performance, and memory bandwidth