When: Pullback to $540-550 (Bollinger mid-band $542, 50 SMA zone $509-$540) with confirmed higher-low and MACD positive

Then: Add 25% of target

When: Above $625

Then: Do not initiate; chase risk is asymmetric given 64.8% extension above 200 SMA

When: Close <$429.56 (July 29 capitulation low)

Then: Opens $338-360 reload zone at full conviction sizing

HALF POSITION - Buy one ATR unit at $607, hold second half in reserve for 565-580 or 540-550 zones; defer SOXS hedge initiation 3-5 sessions; reduce position if close below 10 EMA at $592

• Trailing P/E of 156 with forward consensus EPS at $15.58 makes the PEG of 0.62 dependent on forecasts that haven't proven accurate before
• Stock is 64.8% above 200-day SMA with RSI at 65.52; MACD expansion signals extended momentum, not sustainable momentum
• 30-year rates at 23-year high creates untested regime for AMD's institutional ownership base
• Beta of 2.48 means a 6.5% index move triggers the stop at $509, representing 16% drawdown from entry
• Five consecutive quarters of 50% revenue growth and $2B operating income swing are real but face risk from hyperscaler capex pauses, NVIDIA share recovery, or customer concentration
• Meta deal structure issues $0.01 strike restricted equity tied to deployment milestones; $25.5B goodwill creates mechanical dilution risk
• OpenAI warrants: 160 million at $0.01 strike tied to 6 gigawatts deployment represents contingent equity, not value realized
• Institutional news feed returning empty is not evidence of safety; absence of news means absence of information about customer pauses, export rules, or competitor design wins
• Trader's own plan anticipated pullbacks to 565-580 and 540-550, contradicting aggressive analyst's 'buy before pullback' argument
• SOXS hedge bleeds premium in chop; initiation should wait 3-5 sessions to assess stock behavior
• Recession risk flagged at 8% with credit channels tightening creates correlated downside across AI infrastructure complex
• 10 EMA at $592 is the critical level; a close below signals the pullback character changes from healthy to concerning

→ vs aggressive: Aggressive analyst dismisses trailing P/E of 156 as 'the wrong number' and anchors to forward consensus EPS of $15.58, but consensus is a forecast by analysts who missed the 1,400bps margin expansion being celebrated

→ vs aggressive: Claims 30-year rate at 23-year high means 'this is fine' when multi-decade highs actually mean the regime is untested relative to this valuation

→ vs aggressive: States multiple compression 'already happened' but cannot falsify this claim; 64.8% extension above 200-day SMA suggests more room for mean reversion than expansion

→ vs aggressive: Calls Meta deal dilution defense 'underwriting the compute roadmap' when the balance sheet calls it new shares outstanding; the trader themselves flagged this as warranting respect

→ vs aggressive: Treats 160 million OpenAI warrants at $0.01 strike as 'capital structure innovation' without acknowledging the contingent dilution if 6GW deployment milestones are met

→ vs aggressive: Claims China lobbying is 'not a sign of weakness' when lobbying is what companies do when at risk of losing access

→ vs aggressive: Treats absence of AMD-specific news in 7-14 days as 'no negative catalysts' when institutional news feed returning empty is a flag, not a green light

→ vs aggressive: Argues 'buy before pullback arrives' despite trader explicitly structuring the plan around waiting for pullbacks to 565-580 and 540-550

HALF POSITION INITIATION at 607 with stop at 509; scale into 565-580 and 540-550 zones; defer SOXS hedge; reduce to half size on close below 592

• Forward consensus carries forecast risk and should not be treated as fact; PEG of 0.62 is anchored to a growth rate analysts did not predict a year ago
• Warrant conversion contingent on six gigawatts of deployment creates symmetric risk: upside if milestones are met and dilution if they are not
• Information scarcity (empty institutional news feed) is a risk flag, not a bullish signal; absence of evidence is not evidence of absence
• Lobbying activity indicates real exposure is at risk, not just management defending revenue; both the risk and the defense can be true simultaneously
• Meta restricted equity is both a financing mechanism for compute demand and real dilution risk; the conservative view is too dismissive of strategic value while the aggressive view is too sanguine about dilution
• Current rate regime is largely priced in at 96 percent no-cut probability; surprise hawkishness is the primary risk, not the baseline staying intact
• Technical extension at 64.8 percent above 200-day SMA is extreme; expanding MACD momentum after vertical moves often precedes sharp reversals; shallow pullbacks after vertical moves often resolve into deep corrections
• Initiate half position at 607, stop at 509, hold second half in reserve for 565-580 scale-in zone and 540-550 aggressive add zone if 50 SMA holds
• SOXS hedge should be deferred until 10 EMA at 592 breaks or 30-year yield backs up; hedging into chop bleeds premium through daily rebalancing decay
• Any close below 592 (10 EMA) should trigger reduction to half size, not an add, as that level defines the boundary between healthy pullback and concerning technical event
• Warrants are priced into the multiple partially but not fully; renegotiation or restructuring of milestones would be a meaningful negative
• Data gaps argue for smaller initial size and tighter management, not for paralysis; trade can still be taken at disciplined sizing

→ vs aggressive: Forward P/E is not fact but forecast; PEG of 0.62 is anchored to analyst predictions that did not hold a year ago, and treating consensus as certain is the same error in reverse

→ vs aggressive: Warrant conversion clause requiring six gigawatts of deployment is an aggressive target; the structure creates both upside and dilution if milestones are met, and warrant math going viral is not the same as realized value

→ vs aggressive: Absence of AMD-specific news is a flag, not a green light; information scarcity treats risk as confirmation rather than warning; no data means we cannot rule out paused orders, draft export rules, or competitor design wins

→ vs aggressive: Chasing before the pullback the trader already identified contradicts the trader's own framework; arguing to buy now at the top of a 64.8 percent extension while the plan calls for scaling in on pullbacks is internally inconsistent

→ vs conservative: Meta restricted equity is not free dilution; it is financing compute demand with equity rather than debt, and AMD received a binding compute commitment from a major hyperscaler in return

→ vs conservative: Rate argument that current regime is untested is overstated; long-duration tech has traded through multiple rate regimes, and the 96 percent no-cut probability is already in the tape; surprise hawkishness is the real risk, not the baseline

→ vs conservative: Dilution defense is too one-sided; the strategic value of equity-linked compute deals with the largest hyperscalers creates alignment that should not be dismissed