MU
Micron Technology, Inc.$MUStrong FundamentalsStrong FundamentalsRevenue growing 196% YoY at strong marginsStreet coverage with positive forward estimatesConsistent chatter on X (67.8K/wk), no spike
SemiconductorsBullish sentiment
Micron dominates the memory-cycle narrative after CEO Sanjay Mehrotra announced a raise of planned U.S. investment to over $250B through 2035, with a target of 40% of DRAM production onshore and 90,000+ jobs, alongside a $3B ecosystem investment. Bulls hammer that MU is up roughly 800% YTD yet forward P/E has compressed from 10x to 6.4x on ~345% YoY revenue growth, and that Jensen and Michael Dell are publicly calling memory the new AI bottleneck. A cautious bear thread points to a possible head-and-shoulders, a recent trendline break, and 'top-signal' behaviour from newer retail entrants piling into the trade.
Proven numbers
Micron is the U.S. leader in DRAM and high-bandwidth memory — the specific chip that every AI training and inference cluster is bottlenecked on. The story has shifted from "will orders come" to "how long can this run," and the tape is showing the first signs of digestion.
How the pieces fit:
• Q3 revenue landed at $41.5B and beat estimates by 20%, because HBM shipments track new GPU cluster stand-ups almost one-for-one, and consensus already models revenue nearly tripling to $128B next fiscal year.
• The onshoring commitment adds federal weight to customer trust: management raised planned U.S. investment to $250B through 2035 with 40% of DRAM production onshore, meaning national-security-sensitive orders now have a defensible domestic supply story.
• Valuation has actually compressed: forward P/E has fallen to roughly 12x FY27 EPS as earnings have caught up with the price, so the risk isn't multiple compression, it's the cycle itself rolling over.
• The insider signal cuts the other way: officers sold $66M in the last two weeks against zero open-market buying — after a 633% twelve-month return, employees taking chips off the table is a check, not a confirmation.
The run continues as long as HBM order books keep building and the hyperscalers stay in build mode, and cools further the moment inventories at Nvidia or a major customer show digestion. Position sizing matters more here than direction.
