Burry argues that the rally has reached "historically extreme" levels, with Micron stock more extended above its 200 day moving average than at any point since 1984, "not even during the dot com peak." Burry stated, "Micron defines cyclical like no other," citing 34 drawdowns of more than 30% over 42 years, a median return on invested capital (ROIC) of 4%, and return on equity (ROE) of 7%, which he called "frankly terrible." He added that "one quarter in every three, Micron is a destroyer of capital," with free cash flow negative 48% of the time. I donât know Michael .. this time you just talk with no 13F to prove positions.. I donât like people who just talk. This guy has a history of being way too early on short positions, he saw the sell off and decided to jump in. Why jump in $1058. Currently $975 with a P/E (TTM)22.09 (EPS TTM $44.17) Forward P/E (NTM): 6.80 6.80??? This cyclical talk is a false claim. And the reason being is because each chip Jensen introduces goes with the amount of power the US has as they upgrade the grid as we get closer to nuclear. (Remember the AI factory chart I showed you) Letâs take a look at microns capex spending and investments globally Hereâs the full global expansion map United States ($200B total) Boise, Idaho Idaho 1 leading edge fab, wafer output pulled forward to mid 2027 Idaho 2 wafer production expected late 2028 Clay, New York up to four leading edge fabs, the flagship being a $100B âmegafabâ complex; Fab 1 construction begins late 2026, DRAM production 2029-2030 long term buildout spans two decades Manassas, Virginia expansion/modernization of existing fab plus an HBM packaging/assembly plant (expected online after Idaho HBM ramps, likely assembling HBM5/HBM6 late this decade) Asia-Pacific Hiroshima, Japan the $9.3B/„1.5T expansion just broken ground (your article); HBM chip production, commercial shipments summer 2028 Japanâs government backing „775B total a second Japan fab module is also planned Singapore HBM packaging facility (broke ground early 2025, contributing to production 2027) plus two existing 3D NAND fabs, one of which (Fab 10B) is being expanded Taiwan two existing DRAM fabs (near Taichung and Taoyuan); plus the newly acquired Tongluo site (from PSMC, $1.8B) getting a twin fab, construction starting this summer, shipments from fiscal 2028; original P5 fab ramping DRAM in H2 2027 India & Malaysia back end assembly and test hubs (not fabs) five facilities two Singapore NAND fabs, one Hiroshima DRAM fab, two Taiwan DRAM fabs plus Singapore HBM packaging. The historical âMicron is cyclicalâ argument assumes memory demand follows the old boom bust chip cycle inventory glut, price crash, recovery. But this cycle is gated by something structurally different physical grid capacity and nuclear buildout, which donât oscillate the way semiconductor inventory does. Grid interconnects and reactors donât get âoverbuiltâ and then crash in price the way DRAM fabs historically did. So if GPU/memory demand is paced by watts available rather than by speculative capex cycles, the entire cyclicality framework Burry is using (34 drawdowns over 42 years) is built on a pre AI era pattern that may no longer apply. Memory itself can still overshoot even if power is the gate. If Samsung/SK Hynix/Micron/CXMT all race to add fab capacity because they believe the power buildout locks in years of guaranteed demand, they could still collectively overbuild memory supply relative to how fast gigawatts actually come online the power constraint caps deployment, not necessarily production. A supply/demand mismatch inside a power constrained world is still possible, just on a longer wavelength Executives are saying shortages persist past 2027 Samsungâs chip division just posted record profit specifically because AI driven memory shortages are expected to strain capacity past 2027 thatâs the companyâs own forward guidance, not analyst speculation. South Korea just backstopped this with $520B in state coordinated investment (10x the CHIPS Act) four new fabs, HBM dedicated facilities which only makes sense if the government believes demand visibility extends far enough to justify decade scale infrastructure, not a 2-3 year chip cycle. software engineers, vibe coders etc anyone in this ai technology space will tell you how important memory is. memory makers are not racing ahead of the grid to build speculative capacity that could crash theyâre capacity constrained by their own cleanroom/fab economics (multi year builds) at the same time end demand is capacity constrained by grid interconnects. Two independent bottlenecks (fab lead time + grid lead time) are stacked on top of each other, both multi year, both non cyclical in the traditional inventory glut sense. Thatâs a materially different setup than 2018 or 2022, when memory makers did overbuild into a demand air pocket. Burryâs 34 drawdown dataset is drawn from a world where capacity could outrun demand quickly that mechanism looks structurally weaker right now on both the supply and deployment sides. SKHynix is also coming to the Nasdaq and memory pricing will continue to increase. SK Hynix confirmed and imminent. Board approved issuing up to 17.79M new shares (2.5% dilution) for ADRs, trading on Nasdaq starting July 10, 2026 just days away. Stock surged 10%+ on the announcement and briefly overtook Samsung as Koreaâs most valuable listed company. I believe will see Samsung next. Right After SKHynix now thereâs no timeline on this for Samsung but you can see whats happening here. The value of memory is being rerated right infront of you. Like GPUs once upon a time. However, if we want to look at this on geopolitical standing , The date to look for and Iâve said it before is November. Not now. I want to see you win. Like always my friends Iâll keep you posted. Remember to do your own research, your conviction lays with your confidence . (This is not financial advice.) To my American friends Happy 4th of July! 𧚠Enjoy the rest of your weekend! https://finance.yahoo.com/markets/stocks/articles/michael-burry-shorts-micron-adding-141819112.html
