The SoCAMM2 “cut” headline is being misread. I’m sorry if you think it’s bearish, but it’s not.
$NVDA halved 192GB module orders but 6x’d 96GB and bumped 64GB +50%. Total LPDDR demand into Vera Rubin: +10–20% vs prior forecast.
The mechanism is the tell. Memory supply is tight enough that Jensen would rather ship lower-config systems now than wait for higher-density modules. And the platforms are field-upgradable to 192GB+ later — so there’s a second demand wave baked into the install base.
Net for $MU:
• TAM into NVIDIA data center LPDDR steps up
• Underlying signal = supply tightness, pricing-positive
• Upgrade cycle = forward visibility into next refresh
• Additive to HBM3E/HBM4 tightness, not a substitute
Channel is pulling harder than sell-side has modeled going into 6/24.
Today was a gift dressed as a warning.
$MU closed down 13.25% to $864.01 on the same day Jensen Huang officially named them an HBM4 supplier. Sit with that for a second.
The largest fundamental catalyst the name has had outside of an earnings print, formal confirmation that Micron is the third name in the most critical AI memory supply chain on the planet, and the stock got blowtorched. That isn’t a thesis break. That’s a sector-wide forced de-risking event riding on top of a single macro print, and history says these are exactly the setups that get bought back fast when the next real catalyst arrives.
The damage tape:
$SPX $NDX SOX $NVDA $META $AVGO $AMD $TSLA all hammered
What actually happened: NFP printed 172k vs 80k expected, the 10Y ripped, and the Fed cut narrative got deferred. Not killed. Deferred. One soft CPI or one weaker labor print rebuilds it. That repriced long-duration cash flows lower for a session. Layer that on top of Wednesday’s $AVGO print where the market decided AI capex visibility past 2026 was less clean than consensus, and you get a two-day shakeout of the most-loved names.
Here’s what the bears are missing: the AVGO concern was about custom ASIC mix and timing, not memory demand. If anything, more ASIC content in the AI buildout argues for more HBM, not less. The memory cycle thesis didn’t get a single negative data point this week. It got a positive one when Jensen confirmed MU for HBM4. Price and fundamentals just decoupled, which is the textbook definition of an entry.
The Asia overnight tape tells the same story.. When the entire memory complex moves in lockstep on a U.S. macro print, the move is mechanical, not fundamental. Forced selling has a signature, and that signature is symmetric, indiscriminate, and brief. Real thesis breaks look different.
Where this goes:
The 6/24 $MU print just became the most asymmetric setup of the quarter. IV will keep bidding into the catalyst, which marks vega-heavy structures higher even at lower spot. If the print delivers on memory pricing and HBM guide, and everything from the Jensen confirmation to the AVGO read-through says it will, today’s selloff was the entry of the year, and the squeeze higher pulls in every macro tourist who just got shaken out at the lows.
The bigger picture is the same as it was Wednesday morning. AI capex is structural, the memory supply side is genuinely tight, and the cycle has room to run. Today added a discount, not a problem. The names with the strongest fundamental catalysts in the pipeline are exactly where conviction capital loads up while the macro tourists are still bleeding.
Today wasn’t regime change. It wasn’t even a warning. It was the kind of day that makes year-end P&L for the people who can separate price from thesis.
Freak the fuck out and sell everything. Thesis has absolutely changed and everything has never been more over
$MU $SNDK $SPY $QQQ $NVDA $AVGO
(This is sarcasm for people that can’t tell)