D.A. Davidson just slapped a $3,000 price target on $MU — up from $2,100. That's nearly 200% from current levels.

Gil Luria's thesis: Memory is the bottleneck for AI. More memory = faster models, longer context windows, better output. Supply won't catch demand until 2027-2028.

$MU trades at ~6x FY27 earnings. Target implies 19x multiple. If memory becomes the AI arms race everyone's sleeping on, this could print.

Meanwhile $NVDA keeps getting upgrades. Morgan Stanley at $300. BNP just raised to $345 from $285.

AI infrastructure plays are heating up. Watch the semiconductor thesis closely.