When: Price retests $1,000-$1,015 bear-flag resistance

Then: Execute 15-20% incremental trim before Jackson Hole binary catalyst

When: If $1,000-$1,015 not reachable and price breaks down instead

Then: Execute trim on any retest of 50 SMA at $957-$958

Trim 50% of cyclical exposure now, hold remaining 50% with hard stop at $8.95, trim further into strength toward $1,000-$1,015 resistance, avoid initiating fresh longs at $9.35, reassess after Jackson Hole and next earnings setup

• Directional call to reduce exposure is correct; peer divergence is real (SK Hynix +2.5% vs Micron flat on same news)
• Jackson Hole speech tomorrow is a genuine binary event in a name with 2.21 beta with 88% probability of zero rate cuts already priced in
• 80% operating margin in memory is historically the top of cycle, not the middle
• HBM is not generic DRAM—has qualification cycles, customer lock-in through Nvidia's CUDA ecosystem, and contract structures different from spot-driven 2008/2018 cycles
• Nvidia's $279 billion memory commitment gives pricing visibility through at least mid-2027; renewal cliff is a 12-24 month risk, not immediate
• Stock is priced 60% above its 200-day SMA; multiple expansion mathematically capped in current macro backdrop
• Stop at $8.95 is approximately 1.5 ATR below current price—provides room to breathe in a name that routinely moves $60/day
• MACD histogram peaked at +18 just ten sessions ago and is now under +3—deceleration profile that precedes next leg lower
• Conservative discipline means trimming into strength toward bear flag resistance at $1,000-$1,015, not forced selling into weakness
• Remaining 50% position captures upside scenario if HBM behaves like contracted utility rather than commodity; trimmed half locks in gains from $600-$700 entry levels
• The trader's plan acknowledges cycle risk while keeping skin in the game—preserves ability to reload at better prices if Aggressive is correct

→ vs aggressive: Going to zero or shorting into a name with this much structural HBM optionality is not risk management—it is an unhedged directional bet that the cycle math always wins regardless of contract structure or customer behavior, incompatible with fiduciary responsibilities

→ vs neutral: Hold is not a defensive position—it is the most expensive stance in a name with 2.21 beta sitting below rolling 50-day SMA before a binary macro event, accepting full downside exposure on hawkish Jackson Hole surprise, earnings disappointment, and every cyclical rollover signal without any of the protection a trim would provide

Execute fifty percent reduction into resistance zone (1,000 to 1,015), widen stop to 875 or use tranche-based scaling, establish defined re-entry framework on 957 breakout and 895 breakdown, and redeploy trimmed capital into adjacent AI exposure rather than cash.

• Support fifty percent trim as balanced move but recommend execution into 1,000 to 1,015 resistance rather than current levels
• Widen stop from 895 to 875 or use scaled tranche exit to avoid noise-driven liquidation on hawkish surprises
• Build re-entry framework: add back on confirmed 957 breakout with volume conviction, exit remaining position on 895 breakdown with semis complex
• Pushback on Aggressive: bear flag not confirmed—MACD positive, histogram at 2.87, price in upper Bollinger half; treating undecided tape as confirmed bear case
• Pushback on Aggressive: HBM has multi-quarter qualification cycles unlike prior DRAM cycles; not truly comparable to 2008/2018 bust dynamics
• Pushback on Aggressive: 6x forward P/E could equally reflect contracted visibility through 2027 rather than market skepticism
• Pushback on Conservative: fifty percent trim is compromise position, not principled conservatism; true conservatism would be full hold or smaller trim
• Pushback on Conservative: HBM contracts don't insulate from NAND/conventional DRAM cycles or broader memory market dynamics
• Distinctive point: redeploy trimmed capital into equipment makers or downstream AI beneficiaries with different cycle dynamics rather than cash
• Neutral's position is defensible only if willing to accept short-term volatility for structural exposure; Sell only if cycle has already turned

→ vs aggressive: Bear flag interpretation is not confirmed—MACD positive, histogram 2.87 still in positive territory, price within upper Bollinger envelope; this is undecided tape being treated as confirmed breakdown

→ vs aggressive: HBM has multi-quarter Nvidia CUDA qualification cycles that make supply ramp fundamentally different from prior DRAM flooding events; not comparable to 2008/2018 cycles

→ vs aggressive: 6x forward P/E with 0.14 PEG equally reflects contracted visibility through 2027, not just skepticism; cannot selectively trust price signals

→ vs aggressive: 2028 renewal cliff has materially more uncertainty than admitted; treating it as foregone conclusion is analytically sloppy

→ vs conservative: Fifty percent trim is a compromise position that admits Aggressive's bear case partially correct while refusing full commitment; not principled conservatism

→ vs conservative: HBM contracts extend upside cycles but Conservative ignores that memory cycle dynamics still operate on downside—asymmetric reasoning

→ vs conservative: Hold presented as strawman; Hold is defensible if accepting short-term volatility for structural thesis, not inherently untenable

→ vs conservative: Jackson Hole and Polymarket macro signals have been priced in for weeks; positioning around known binary events is typically a wash