Before Micron’s June 24, 2026 FQ3 earnings release, Taiwan-listed memory and storage-linked companies had already laid out much of the cycle shift in public monthly revenue filings. The source argues that Taiwan’s reporting rules, which require listed semiconductor firms to disclose the previous month’s revenue by the 10th, create a timing gap versus Micron’s quarterly reporting schedule. That gap gave investors an earlier read on the DRAM and NAND recovery.
Why Taiwan filings became an early read on Micron
The setup is straightforward. Nanya and Winbond track traditional and specialty DRAM. Phison reflects NAND controller and SSD demand, while Adata is highly sensitive to shortages and price increases as a module maker. According to the source, DRAM and NAND made up all of Micron’s FQ2 revenue mix, with DRAM at 79% and NAND at 21%. That made Taiwan monthly revenue data a useful public proxy for most of Micron’s core business cycle.
There are clear blind spots. Taiwan monthly filings do not capture Micron’s HBM business, because HBM is produced in-house by Micron, Samsung and SK hynix rather than outsourced to Taiwan. The same framework also does not extend to HDD-focused names such as Western Digital and Seagate. So the signal works best as a read on the broader DRAM and NAND environment, not as a complete predictor of Micron’s post-earnings stock reaction.
Signals were already visible in late 2025
The source traces the turn back to October 2025. Nanya’s monthly revenue growth accelerated from 158% year over year in September 2025 to 262% in October, reaching a 50-month high. In November and December, both Nanya and Phison posted record monthly revenue. The article also cites comments from Phison CEO K.S. Pua in early January 2026, saying CSP demand was pushing NAND prices higher and tightening supply.
By the time January 2026 revenue was disclosed, the move had broadened sharply: Nanya was up 608% year over year, Adata 199%, and Phison 189%. All of those figures were public before the decision point discussed in the article: whether buying Micron in March 2026 made sense. The source’s retrospective example says a purchase on any trading day in March and a hold through June 23 would have shown gains of roughly 128% to 155%.
The rally changed character in 2026
The source separates the move into two phases. In the second half of 2025, the rise was driven by a classic memory-cycle rebound and supply-side disruptions, including Samsung’s suspension of DDR5 contract pricing and capacity shifts away from DDR4 in China. That pushed fourth-quarter traditional DRAM contract prices higher by 18% to 23% quarter over quarter.
In 2026, the driver changed. The article says AI demand and HBM allocation started squeezing conventional DRAM and NAND supply. HBM was described as consuming about 20% of DRAM wafer capacity, helping drive a 90% to 95% quarter-over-quarter jump in traditional DRAM contract prices in the first quarter of 2026. It also cites Micron CEO Mehrotra as saying non-HBM gross margin had moved above HBM gross margin, which the article links to Micron’s gross margin approaching 81%.
What Taiwan data could not answer before earnings night
For the June 24, 2026 earnings release, the source says the market was focused less on the reported FQ3 numbers and more on fourth-quarter guidance, HBM pricing, and visibility into 2027. The article lists Micron’s own revenue outlook at about $33.5 billion and gross margin around 81%, while market consensus had already moved higher to roughly $34.5 billion to $35.6 billion in revenue and $19.7 to $20.6 in EPS. Expectations for the next quarter were even higher, at $38 billion to $40 billion in revenue, while options markets implied a 14% to 17% one-day move.
That is where the limits of the Taiwan revenue signal show up. It can point to an early turn in DRAM and NAND demand and pricing. It cannot measure HBM directly, and it cannot replace Micron’s own guidance. In the source’s framing, it works as a compass for the cycle, not a crystal ball for the earnings reaction.