Micron (MU) Stock Jumps on China-Memory Policy and a Bet That the Cycle Has Changed

Why Micron stock jumped: a policy tailwind and a bet that memory has stopped being cyclical

On Monday, August 17, 2026, Micron Technology (NASDAQ: MU) jumped about 5.3% to roughly $1,023 intraday, from a $971.66 prior close, joining a broad memory-and-storage rally [1]. The immediate spark was policy: over the weekend, U.S. Commerce Secretary Howard Lutnick told The Wall Street Journal that the administration "does not approve" of Apple turning to Chinese memory, saying the shortage "has to be solved another way, but not by having America's premier companies use memory made in mainland China" — a stance that reduces a potential Chinese sourcing avenue for Apple (which could still proceed) and is supportive for non-Chinese suppliers including Micron [2]. That landed on top of Friday's provocative call from New Street Research, whose Pierre Ferragu had upgraded Micron to Buy with a $1,250 target, arguing that memory has broken out of its notorious boom-and-bust cycle [3]. Micron is now a $1.16 trillion company, up more than tenfold from its April 2025 lows [1].

Micron is the largest U.S. memory maker — DRAM and NAND flash — and, alongside South Korea's SK Hynix and Samsung, a key supplier of the high-bandwidth memory (HBM) stacked next to AI accelerators. The policy news gave the whole group a same-day lift, but the reason Micron keeps climbing is a bigger debate, the one that has defined memory investing for decades: is this a durable, structurally different business, or a cyclical peak that will eventually mean-revert? The clearest expression of that debate is the analysts' own price targets, which fan out across an exceptionally wide range.

The bull case: why an analyst says the cycle has changed

Ferragu's upgrade is not a momentum call; it is an argument that the structure of the memory industry has changed, with HBM at the center [3][4]:

• HBM soaks up capacity. Current-generation HBM (HBM3E) uses roughly three times the wafer capacity per bit of conventional DDR5 DRAM and requires advanced packaging. As AI demand pulls HBM production higher, it consumes wafers that would otherwise make commodity DRAM — which constrains non-HBM supply and, the argument goes, prolongs the shortage beyond a normal cycle [3].
• A shallower cycle. The core claim is that Micron can stay highly cash-generative even through a milder multi-year downturn, so the amplitude of the memory cycle is shrinking — memory becoming less of a boom-and-bust commodity and more of a structurally profitable business [3].
• Price versus production value. Micron's stock has risen more than tenfold since its April 2025 low while its cost of goods sold is up only about 25% over the same span — which the bull case reads as pricing and mix driving durable value creation rather than a purely cyclical bounce [3].
• A bold long-term scenario. In the most aggressive version, Ferragu sketches peak figures of roughly $150 billion of annual free cash flow and $600 billion of cash by 2030, consistent with a $2–3 trillion valuation — explicitly a long-term scenario if everything lines up, with a materially milder downcycle assumed thereafter, not a base case [3].

The trailing numbers are, for now, spectacular: twelve-month revenue of about $90 billion (up ~167% year over year) and net income near $50 billion [1]. The question is whether those results are a new baseline or a peak.

The bear case is hiding in the valuation

Here is the tension at the heart of Micron today: despite a more-than-tenfold run, the stock trades at only about 7x forward earnings (versus roughly 23x trailing) [1]. That is a striking multiple, but it is important not to over-read it. Consensus earnings estimates are still rising — S&P Global data show adjusted EPS climbing from roughly $73 in fiscal 2026 to about $155 in fiscal 2027 — so the low forward multiple does not mean the market expects earnings to fall next year [1]. What it does reflect is that investors are discounting the durability of those forecast earnings: after decades of memory being a boom-and-bust commodity, the market is unwilling to pay a high multiple for exceptionally elevated forecast profits whose through-cycle durability it doubts. New Street's argument is precisely that this caution is a mispricing — that if the cycle is genuinely shallower, memory earnings deserve a higher, more stable multiple than the market is granting. The opposing view, which we laid out when Micron sold off earlier this summer on DRAM peak-margin concerns, is that memory has looked structurally different at the top of every cycle and always mean-reverted. The low forward multiple is the market keeping that skepticism alive.

Why it matters

Micron is one of the clearest large-cap bets on whether AI has structurally changed memory economics. If the bulls are right, a business the market has long valued at low multiples deserves a structural re-rating, and the direct memory suppliers — SanDisk in NAND, plus SK Hynix and Samsung — re-rate with it. If the skeptics are right, today's exceptionally elevated forecast profits eventually normalize sharply, and the low forward multiple is the market's caution rather than an opportunity missed. The weekend policy news adds a second, more concrete tailwind: reducing a potential Chinese sourcing avenue for Apple is supportive for non-Chinese suppliers like Micron. The stakes are unusually high because the stock has already moved so far — a $1.16 trillion valuation leaves little room for the thesis to be wrong.

How the peers and sector traded the same day

Micron's jump was part of a broad rally across memory and storage on August 17, which tells you the move was about the sector-wide policy and demand narrative, not a single-stock catalyst [6]:

The direct memory read-through runs through SanDisk (NAND) and the Korean DRAM makers SK Hynix and Samsung; Western Digital rallied in sympathy but is a storage, not a memory, play after the SanDisk spin-off. The whole complex moving together on a supply-and-policy narrative is what you would expect if the market is re-rating the AI-memory trade as a group, with the New Street upgrade an extra tailwind specific to Micron [6].

What the Street did with it

The analyst reaction has been overwhelmingly bullish, even as the targets fan out. Friday's New Street upgrade to Buy carried a $1,250 target (about 32% above the pre-upgrade price) [3]. The Street-high sits even higher: UBS's $1,625 target, set back in May by analyst Timothy Arcuri (a more-than-tripling from $535) and maintained since [7]. Across roughly 46 analysts the stock holds a consensus "Strong Buy" with an average target near $1,502 — but a low estimate of $361 and a high of $2,200 [5]. That dispersion, more than any single target, is the useful signal: the sell side is broadly bullish on near-term demand but deeply split on how much of the current earnings power is durable.

What to watch

• HBM supply and pricing. The bull thesis rests on HBM staying structurally short. Watch capacity announcements from Micron, SK Hynix and Samsung, and any sign that HBM supply is catching up to demand [3].
• Margins through a soft patch. The "shallower cycle" claim will be tested the first time DRAM pricing weakens; watch whether Micron's gross margin and cash flow hold up far better than in past downturns [3].
• China policy follow-through. The weekend signal discouraging Apple from Chinese memory is supportive for non-Chinese suppliers like Micron; watch whether it hardens into broader, formal restrictions or fades [2].
• The forward multiple. A re-rating of the ~7x forward P/E toward a higher, mid-cycle multiple would be the market conceding the durability case; a multiple that stays low even as earnings rise says the skepticism persists. Moves are tracked on the MU stock page and the earnings calendar [1].

Illustrative valuation sensitivity

The scenarios below are a descriptive, author-weighted exercise — not a forecast, target, recommendation, or intrinsic fair-value calculation. The upside and middle cases sit near published analyst targets (which are 12-month), while the downside is the author's own cyclical case; note that those 12-month targets and the structural bull thesis (centered on 2030 economics) are different horizons, so this is a rough sensitivity, not a precise model [5]. The weights are the author's subjective assumptions. The percentages are illustrative and sum to 100%.

Weighting those (0.30 × $1,625 + 0.40 × $1,250 + 0.30 × $500) gives an author-weighted scenario value near $1,138, close to the ~$1,023 quote [1]. But the point of Micron is not a point estimate — it is the width of the band, from around $500 to above $1,600. Few mega-caps carry that much unresolved disagreement about whether their current earnings are a floor or a ceiling. This is descriptive analysis of a corporate event, not investment advice.

MU data snapshot — August 17, 2026 (intraday, session open)

The August 17 figures are intraday snapshots that moved through the session, not closing prices; the multiples are quoted at the current price rather than Friday's close.

The setup at a glance