Two things happened simultaneously on the evening of May 26.
Xiaomi released its first-quarter 2026 financial report. Total revenue was ¥99.1 billion, a year-on-year decrease of 10.9%; adjusted net profit was ¥6.07 billion, a staggering 43.1% drop year-on-year. The revenue from the smartphone business was ¥44.3 billion, down 12.5% year-on-year, with a gross margin of 10.1%, a 2.3 percentage point decrease from the same period last year.
During the financial report conference call, Xiaomi Group's President Lu Weibing mentioned a number: compared to the same period last year, the price of the same version of memory has skyrocketed nearly fourfold. For a smartphone with a configuration of 12GB LPDDR5 + 512GB UFS, the memory cost alone has increased by about ¥1500.
He said Xiaomi "will not pass on the increased memory cost to consumers," but at the same time predicted that the price hike cycle will continue until 2027 or even 2028. In order to survive, Xiaomi proactively eliminated its entry-level models, and its quarterly shipment volume dropped to 33.8 million units.
The second thing, Micron Technology surged more than 19% in a single day, surpassing a $1 trillion market capitalization. UBS raised Micron's target price directly from $535 to $1625, a one-time increase of about 204%, becoming the highest target price among the 46 brokerages covering Micron.
A few days ago, Citigroup had just raised Micron's target price from $425 to $840, and HSBC also increased it from $750 to $1100. Wall Street has not been so unanimously positive on the same cyclical stock for a long time. Micron's stock price was less than $110 a year ago. It has increased eightfold in one year.
On the same day, there was a trillion-dollar celebration for memory sellers and a halving of profits for memory buyers.
Goldman Sachs played an interesting role in this celebration. In December 2025, Goldman Sachs gave Micron a Neutral rating with a target price of $205. In the first quarter of 2026, Goldman Sachs reduced its Micron holdings by nearly 20%.
On March 19, Micron's earnings report day, Goldman Sachs raised the target price from $360 to $400, but maintained a Neutral rating, even though the stock price had already far exceeded $400. Then Micron surged 40% in a week, and Goldman Sachs made a precise empty-handed bet.
On May 17, Goldman Sachs released a storage industry report, concluding it was the "most severe supply shortage in 15 years" and upgraded the overall storage industry rating. However, it still maintained a Neutral rating on Micron with a target price of $400. This eccentric behavior by Goldman Sachs is either the last sober person in this celebration or the one who made the most disastrous empty-handed bet.
However, this strong division is also worth serious consideration.
01 Why the Craze, a New Story Called LTA?
A research report by UBS analyst Timothy Arcuri on May 26th, the core argument is that Long-Term Agreements (LTAs) are fundamentally eradicating the semiconductor industry's cyclicality.
Storage chips are the most commodity-like variety in the semiconductor industry. The prices of DRAM and NAND have followed a brutal rule for forty years, rising for two years, falling for two years, and price collapses have never been absent. The profits of Micron, Samsung, and SK Hynix are like electrocardiograms, and the market has never dared to value these companies based on "steady-state earnings." Over the past forty years, the rough valuation fluctuation range of cyclical stocks is between 8 and 15 times the price-earnings ratio.
UBS's story is that the "cyclical curse" of these companies will be broken, and the protagonist behind it is "AI."
Cloud providers such as Microsoft, Google, Amazon, and Meta have started to proactively sign 3 to 5-year fixed-price long-term contracts with storage manufacturers to secure the supply of HBM and DDR5 in the AI arms race, and these contracts come with prepayments.
These contracts are not the traditional "intentional" agreements in the semiconductor industry; they are binding procurement commitments that lock volume, price, and even wafer capacity.
Microsoft and Google were reported in April to be negotiating a three-year DRAM agreement with SK Hynix, including a prepayment structure. Previously, it was manufacturers seeking orders from customers; now it is customers paying deposits to lock in production capacity. The power dynamics of the supply chain have reversed.
According to UBS's model, if LTA is included in Micron's earnings forecast, even if DRAM spot prices drop by 50% in the fiscal year 2029, Micron's full-year EPS can still be maintained at above $100. LTA can narrow the price fluctuation of DDR from peak to trough by about 50%.
By 2027, 20% to 30% of the industry's total DDR bit shipments will be locked in at fixed prices through long-term agreements. In the DDR5 procurement of leading hyperscalers, 60% to 70% may already be under fixed contracts.
From a valuation perspective, if cyclicality disappears, storage stocks should no longer be valued as cyclical stocks but should be valued as infrastructure utilities, with P/E ratios jumping from 8 to 15 times to 20 to 30 times.
In mid-May, JPMorgan also issued a research report with a similar conclusion, directly titled "LTA Is Eliminating the Storage Industry's Cyclicality." Citigroup's logic is that HBM production will squeeze ordinary DRAM wafer capacity, causing a long-term shortage in general-purpose storage as well.
Micron's soaring stock price is experiencing a Davis double play of profit and valuation system shift.
02 This Storage Is Not That Storage
Wall Street's "storage supercycle" narrative refers to a unified bull market story. However, "storage" is not the same as "storage" at all.
The storage market in 2026 presents three layers of differentiation.
The first layer is AI storage: HBM, server DDR5, enterprise SSD. Prices here are rising, shortages are occurring, and long-term agreements are locking in capacity simultaneously. TrendForce predicts that in the second quarter of 2026, DRAM contract prices will increase by 58% to 63% compared to the previous quarter, and NAND Flash contract prices will rise by 70% to 75%; KnightShield has also stated publicly that capacity in 2026 has been essentially sold out. This layer is the story of Micron's trillion-dollar market capitalization.
The second layer is mobile and embedded storage: mobile DRAM and smartphone NAND. Prices are also rising sharply here. Counterpoint data shows that in the first quarter of 2026, DRAM prices increased by over 50% compared to the previous quarter, and NAND Flash prices increased by over 90%.
According to relevant TrendForce reports, memory typically represented about 10% to 15% of a phone's bill of materials (BOM) in the past and has now risen to 30% to 40%, with more pronounced pressure on low-end models.
Xiaomi is right in the middle of this. Its pain is that "AI has taken up production capacity, leaving less for smartphones, forcing smartphone manufacturers to pay a higher price for the remaining capacity."
The OEMs have given production priority to AI customers, leaving smartphone manufacturers with limited options for contract procurement. If you want to ship goods, you have to buy at the new contract price; if you don’t buy, both the production line and the new product schedule will be affected.
The third layer is the PC retail spot market: DDR5 modules, consumer-grade SSDs. Here, there has been a reverse fluctuation. A report from TrendForce shows that by the end of March, the price of a 32GB DDR5 module in the Chinese channel dropped from nearly 3000 RMB to between 500 and 1050 RMB, with some clearance prices as low as 1950 RMB; Tom’s Hardware also notes that some DDR5 products in China and overseas retail markets have fallen back 25% to 30% from their peak.
However, this is mainly a split between retail spot market and contract procurement. PC channels have inventory and can sell off excess stock; smartphone manufacturers have to adhere to contract purchases, without the option to sell off excess stock.
Within the same "storage" industry are three layers with different directions. The essence of this differentiation is that the three storage giants are shifting wafer capacity from consumer-grade to AI. HBM production is taking over regular DRAM wafers, enterprise SSDs are taking over consumer-grade NAND supply, leaving less capacity for smartphones and PCs. Smartphone manufacturers are forced to accept price increases to meet shipment demands, while PC channels can sell at discounted prices due to sufficient inventory.
Micron and others have actively chosen to allocate capacity to AI clients who are willing to pay more. In the short term, this is a smart product structure upgrade. However, it also means that Micron is closing off its retreat. If AI demand slows down, the capacity may not easily switch back.
Micron's financial report shows that, compared to the previous quarter, DRAM bit shipments only grew at a mid-single-digit rate, NAND bit shipments only grew at a low-single-digit rate, with the growth mainly driven by an increase in ASP. Micron's story today is solely focused on the "extreme shortage in AI storage."
Micron has bet everything on this segment.
03 Can Long-Term Contracts Really Eliminate Cycles?
The logic behind long-term contracts seems solid. With the pace of AI spending, the supply elasticity of storage chips is very low, HBM capacity takes 18 to 24 months from planning to production, and producing HBM will take away capacity from commodity DRAM wafers. Cloud providers sign long-term contracts because they're concerned about "AI project delays."
Different institutions have different methodologies for calculating AI CapEx, but the direction is consistent: AI infrastructure investment is transitioning from the level of hundreds of billions of dollars to nearly a trillion dollars. According to some market models, this represents an annualized capital expenditure growth curve of close to 40% to 50%.
However, there is nothing in the physical world that grows at over 40% indefinitely. It doesn't require an AI bubble to burst; a reduction in growth from 45% to 20% is enough to potentially reverse the supply-demand balance of storage chips within 18 months. The three storage manufacturers are currently ramping up production like crazy, with Micron's CapEx for FY2026 at $25 billion and an additional $10 billion planned for FY2027.
There is also one thing that cannot be ignored: when a company's revenue growth depends entirely on price elasticity rather than volume elasticity, the story becomes fragile. Micron's shipment volume only increased by 4% to 6%, and the revenue growth of 196% was mainly driven by price increases. Prices can rise and fall, and the decline is much faster than the rise. This is also the essence of cycles.
Let's do a simple math problem.
Micron's current market capitalization is $1 trillion. Micron has raised its CapEx for FY2026 to over $25 billion and expects a significant increase in capital expenditures for FY2027, with some market reports suggesting an additional increase of over $10 billion.
Micron's non-GAAP net profit for Q2 of FY2026 is approximately $14 billion, simple annualized to around $56 billion, corresponding to about an 18x P/E ratio. If subsequent price increases and long-term contracts continue to be extrapolated, the P/E ratio can be reduced to around 15x.
It may seem "cheap." But the denominator of this P/E ratio is a DDR4 contract price that has increased tenfold in 15 months, an HBM sold out for the whole year, and a super-cycle peak profit where the gross margin has surged from 36% to 75%.
Multiplying earnings at the peak of a cycle by a seemingly "reasonable" multiple to arrive at a seemingly "not expensive" valuation is the classic valuation trap seen at market tops.
In the year 2000, Cisco traded at a "mere" PE multiple of over 60, built on the foundation of 15 consecutive quarters of revenue growth exceeding 50%. When the growth rate dwindled from 50% to 20% and then to 0%, the stock price could plummet by 80% even without a significant drop in earnings, as both the multiple and earnings contracted simultaneously.
From the Davis Double-Click to the Double Whammy.
History tells us one thing: in the commodity markets, long-term contracts have never been a one-sided "floor." They protect the buyer in an upswing and the seller in a downturn, provided both parties have the ability and willingness to fulfill. The moment when long-term contracts are truly needed is precisely when they are most likely to fail.
This is not to say that Micron is definitely in a bubble. The demand for AI in computing power and storage may indeed be structural, LTA may really be rewriting industry rules, and a trillion-dollar market cap may just be the starting point.
But when the whole Wall Street chorus simultaneously chants "this time is different," it is at least worth pausing to ask: the last time everyone was so certain, what happened next?
In a sense, making money often requires embracing the euphoria of a bubble.
However, it took Cisco about 25 years to surpass the closing high set during the dot-com bubble, arriving at today's AI era, where the internet has indeed changed everything.
Welcome to join the official BlockBeats community:
Telegram Subscription Group: https://t.me/theblockbeats
Telegram Discussion Group: https://t.me/BlockBeats_App
Official Twitter Account: https://twitter.com/BlockBeatsAsia
