J.P. Morgan: after accounting for de-spec, expect HBM S/D shortage of -20%/-19%/-16% in CY26E/27E/28E, with accumulated shortage widening to 23 weeks
by CY28E.
$MU$DRAM$EWY$DISK
we put 80% probability on $MU Q4 revenue > $52B when it reports on Wed
micron guided $49Bโ$51B.
analysts: highest ~$52B
our base case is $56.2B.
the big question is how much of higher memory prices reaches reported revenue?
full note: tessara.tech/views/micron-q4โฆ
$MU Micron just fit 4 houses into 1 apartment building...
Old way to get 512GB of server memory: 4 X 128GB sticks, 4 slots, 44 watts.
New way: 1 X 512GB DDR5 stick, 1 slot, 16 watts.
Same land, 4x the people, 60% less power.
When every slot gets the upgrade, one server holds 12TB.
This is how AI stops starving for memory.....
$MU | Baird ๐บ๐ฎ๐ถ๐ป๐๐ฎ๐ถ๐ป๐ ๐ข๐๐๐ฝ๐ฒ๐ฟ๐ณ๐ผ๐ฟ๐บ on ๐ ๐ถ๐ฐ๐ฟ๐ผ๐ป ๐ง๐ฒ๐ฐ๐ต๐ป๐ผ๐น๐ผ๐ด๐, raises PT to $๐ญ๐ฑ๐ฎ๐ฌ from $๐ญ๐ฎ๐ด๐ฌ
Analyst sees stronger AI demand, slower 2027 DRAM supply growth and higher HBM margins supporting Micron.
>Non-stop tech call buying since last week
>Tens of billions in dark pool prints stacking underneath
>Positive Gamma Distribution on All-Expirys
This is what positioning looks like before a GIANT move.
INNIO says grid connection times have stretched from ~2 years historically to 7+ years today, which is why hyperscalers are now signing 15-year contracts for BTM power.
$MU Wells Fargo just cut its price target $1,525 to $1,400 while raising FY27 EPS about 5% to $166 and FY28 more than 10% to $178. Overweight kept. That is a multiple call, not a business call: Rakers now models $261.3B revenue and $125B plus annual free cash flow in FY27, roughly 5% above Street, with 100% shareholder returns starting Dec 2026.
He is lowering what he thinks the market will pay, not what Micron will earn, days before the Sept 30 print.
$SPCX | Mizuho ๐ฟ๐ฒ๐ถ๐๐ฒ๐ฟ๐ฎ๐๐ฒ๐ ๐ข๐๐๐ฝ๐ฒ๐ฟ๐ณ๐ผ๐ฟ๐บ on ๐ฆ๐ฝ๐ฎ๐ฐ๐ฒ๐ซ, PT $๐ฎ๐ฌ๐ฌ
Analyst sees stronger confidence in SpaceX sustaining premium compute pricing into 2027 as incremental capacity comes to market.