To The Moon

$Direxion Daily MSCI South Korea Bull 3x Shares(KORU)$ This should run July 10th with SK ipo. $Roundhill Memory ETF(DRAM)$ 

$Micron Technology(MU)$ $Roundhill Memory ETF(DRAM)$ $Invesco QQQ(QQQ)$ $SPDR S&P 500 ETF Trust(SPY)$ $NVIDIA(NVDA)$ Nvidia used to be called "just a gaming GPU company." Then AI rewrote the demand structure, and a $5 trillion market cap followed. Now Micron is called "just a memory company." AI and HBM are rewriting the demand structure right now. You know what follows. Same denial, same opportunity.

$Micron Technology(MU)$ $Roundhill Memory ETF(DRAM)$Lenovo reports that memory prices are skyrocketing and are set to persist as the 'new normal' until after 2030.

$Roundhill Memory ETF(DRAM)$ SK Hynix is coming to America in a few weeks. I'm sticking with DRAM plays in case some people move their MU profits to this 50% supplier of HBM RAM.

$Roundhill Memory ETF(DRAM)$  Just a bit annoying that it was above $81 before Korea got all worked up. Then Micron unleashes one of the most epic earnings reports in stock market history, and now we're hovering well below $81. Gotta love the markets.

$Micron Technology(MU)$ The AI infrastructure shift is slowly pulling DRAM into the picture. That line about "Agentic AI is structurally reshaping data center infrastructure" basically means we're moving beyond just GPU-heavy racks. The architecture is becoming more layered now: CPU racks handling agent control and execution logic; Storage racks scaling for persistent context and memory-heavy workloads. What I find interesting is how AI is evolving from "single-shot compute" into something much more stateful and continuous. If that trend holds, $Roundhill Memory ETF(DRAM)$  demand isn't just tied to cycles anymore—it starts behaving more like core infrastructure demand across multiple layers. Feel

Micron is getting a lot more demand visibility through these DRAM take-or-pay contracts. It feels like memory is becoming less purely cyclical and more locked-in at scale. The key idea is: Customers commit to fixed volumes over multiple years. The take-or-pay structure reduces demand uncertainty. It signals a tight supply-demand balance in AI memory. Why this matters: Revenue becomes more predictable. Downside risk in cycles is partially cushioned. AI demand is strong enough to justify these long-term contracts. In simple terms: This is no longer just about spot market memory pricing. It's increasingly contract-driven demand. From this angle, $Micron Technology(MU)$  feels closer to a structured demand business than a pure cyclical chip name.

$Micron Technology(MU)$ $Roundhill Memory ETF(DRAM)$ Bears probably won't like this one. The forward setup is starting to price something very different from the old memory cycle narrative. Modeling ~85% gross margins into 2027 tells you the market is no longer thinking in "typical semiconductor cycle" terms—it's pricing AI-driven scarcity, pricing power, and tight supply dynamics. When the margin structure shifts like that, the debate stops being about near-term noise and starts becoming about whether the old cycle framework even applies anymore. Either the model is wrong, or the cycle has fundamentally changed.

$Roundhill Memory ETF(DRAM)$  I read that there's still a three-year gap in memory chip production capacity. When you get a nice dip like this, it seems like a solid opportunity to me.

$Roundhill Memory ETF(DRAM)$ This price action is shaking off the weak hands. Institutions are accumulating because they know MU's earnings report is going to be extremely strong.