AMD

Convergence bar not cleared — win-probability 58 is below the 65 threshold for a new buy despite decent thesis and catalyst. Extreme valuation and weak momentum are structural headwinds. Interesting but not actionable from here.

Exceptional catalyst day with Advancing AI 2026, ROCm launch, and multiple chip-race narratives. But win-probability at 58 misses the 65 new-buy bar and risk quality is poor at 25. Great story, uncertain outcome — convergence rule prevents new_buy. Catalyst adds urgency but not direction when execution risk is this high.

Loudest catalyst day in the batch and strong own-momentum, but Win-Prob at 58 falls below the 65 convergence bar for new buys, and thesis_pct 68 is only marginally above 60. Not a convergence signal. Stretched valuation and poor risk quality argue for patience. Skip — catalyst provides urgency but not direction on a name that hasn't converged.

Transformative Helios/MI455X/Anthropic launch day is a genuine catalyst and momentum is roaring, but win-probability at 58 sits below the 65 convergence bar for new_buy and valuation is stretched. Thesis compelling, execution risk real. Watch for a pullback or a win-prob upgrade before initiating.

Big catalyst day with the $5B Anthropic partnership and MI450 news, but win-probability only 58 and thesis at 68 doesn't clear the 60/65 convergence bar for a new buy. Momentum percentile 96 is exactly the extension pattern that burned us on MRVL and ARM previously — strong buy signals near highs have a poor track record. Skip on convergence failure and extension risk.

Strong catalyst convergence today with Helios AI Rack and Microsoft partnership, and momentum is red-hot at 96 — but that's exactly the pattern our recent-losers learning flags (extension near highs is a stronger sell signal than momentum is a buy). Win-probability at 58 doesn't clear the 65 convergence bar for new_buy. Skip.

Catalyst is real — Helios with Microsoft/Anthropic is material — but momentum at 88 puts it in the extended zone that has repeatedly punished us (MRVL, ARM lessons). Win-probability below the 65 convergence bar for new_buy. Valuation stretched. Catalyst drives urgency, not direction, and here the direction check fails.

Bull thesis at 68 clears one convergence bar but Win-Prob at 58 fails the 65 threshold for new_buy. Intrinsic quality weak (RQ 25) on stretched valuation. Momentum elevated. Not a clean convergence — skip.

Watchlist name with decent catalyst and win-prob but thesis at 68 combined with poor risk quality and price extended near 52wk high triggers exactly the MRVL failure pattern from our learnings. Convergence isn't clean — thesis below the strong-thesis threshold. Skip until either a pullback or a thesis upgrade.

Constructive bull setup with MI300X/MI325X ramp but fails the convergence bar (thesis 68, need 60+; win-prob 68, need 65+ — technically clears but barely). Low risk_quality of 25 combined with proximity to 52wk high echoes the MRVL 'bought at highs with poor risk quality' loss pattern. Not compelling enough as a fresh buy.

Fails new_buy convergence bar (thesis_pct=68 < 60 threshold met but win_prob=68 just above 65). Strong catalyst and improving GPU narrative, but risk_quality of 25 combined with stock near 52wk high echoes the MRVL loser pattern — extended entry with poor risk quality. Watchlist-worthy but not a buy here.

Constructive thesis and decent win-probability, but fails the new_buy convergence bar (thesis 68, win_prob 68 — neither clears 60/65 combo cleanly at the level I want for a fresh entry into a name with risk_quality of 25 and 185x PE). The MRVL lesson — buying near 52wk high with poor risk_quality — applies directly. Watchlist only.

Both thesis and win-prob just above the new-buy bar but neither is convincing, and risk_quality is poor at 25 with an extreme valuation. Pattern learning explicitly flags 'stock near 52wk high with low risk_quality' as a size-down/skip signal (MRVL lesson). Constructive but not a convergence buy.

Constructive setup with near-breakout momentum and upcoming positive catalyst, but thesis_pct at 68 falls short of the 60+ AND win_prob 65+ convergence bar on the win_prob side barely — and MRVL learning warns against buying at 52-week highs with low risk_quality. Skip on convergence discipline; would reconsider on pullback or post-catalyst confirmation.

Bull thesis is real but thesis_pct (68) sits just above the 60 bar while all other cross-currents — extreme valuation, low intrinsic risk quality, and price extended near 52-week high — echo the MRVL loss pattern. Fails the convergence bar for new_buy on quality. Constructive lean but not enough to initiate.

Momentum and catalyst are strong but thesis (68) and win-probability (60) don't clear the 60/65 convergence bar for a new buy. Extreme valuation and poor risk quality echo the CRWD-type setup — chasing a high-PE AI name near highs without conviction convergence. Constructive lean, not a buy.

Fails the new_buy convergence bar (win_prob 60 < 65). Momentum is fine but valuation is stretched and risk_quality is weak. Nothing here says we need to add AI-chip exposure at these levels when we already own NVDA/AVGO/MRVL/TSM within the fund. Watchlist skip.

Strong momentum and today's direct AI data-center catalyst are supportive, but win_probability at 60 fails the 65 convergence bar for a new buy. At 172x PE with intrinsic risk_quality of 25, the CRWD-style trap (great narrative, no valuation discipline) is exactly what we should avoid. Skip on convergence rule.

Strong momentum but win_probability=60 and thesis_pct=68 both miss the new_buy convergence bar. Extended multiple with recent -4% pullback reads like late-cycle chase, not durable setup. Skip.

Win-probability at 60 fails the 65 convergence bar for new_buy. Positive catalyst and decent momentum, but 172x PE and low intrinsic risk quality echo the CRWD pattern — narrative without valuation discipline. Constructive stock, but not a fresh entry here.

AMD has the best momentum in the batch and rides the 'NVDA alternative' narrative, but at 172x PE with only a 60 win-probability the risk/reward is asymmetric the wrong way. Fails the new_buy convergence test (needs both thesis>=60 AND win_prob>=65). Skip on forward view.

Constructive setup with strong momentum and improving competitive position, but win_probability of 60 falls below the 65 threshold required for a new_buy. 172x PE and low risk_quality mean the margin for error is thin. Watchlist status appropriate; would need either a valuation reset or clearer win-probability signal to initiate.

Bull setup is decent but does not clear the 60/65 convergence bar (win_prob 60). Extreme PE plus low intrinsic risk quality make this a wait-and-see. Skip on watchlist.

Best momentum in the batch and a directly relevant agentic-AI catalyst, but win-probability of 60 fails the 65 convergence bar and risk quality is poor. Constructive story but doesn't clear the new-buy threshold.

Near 52w high with decent momentum and thesis, but win_probability of 60 fails the 65 convergence bar for new_buy. PE 172 and risk_quality of 25 are exactly the warning signs we ignored on AVGO/IREN. Skip — no forced entry without convergence.

Momentum is screaming and the breakout-at-highs pattern matches our winning setup, but win_probability of 60 is below the 65 convergence bar for new buys and thesis at 70 is just at the threshold. Per our learnings, breakouts at extreme PE without strong intrinsic quality are exactly the Broadcom/IREN trap. Constructive but not a convergence buy.

Momentum and proximity to highs are attractive and fit the winning breakout pattern, but win_probability of 60 sits just below the 65 convergence bar for a new buy and risk_quality is poor at 25. Without a named hyperscaler/Nvidia partnership headline specific to AMD, this isn't the setup that's worked. Skip.

Bull thesis with real momentum and a concrete analyst catalyst (Morgan Stanley EPYC Venice unit projections). However win-probability of 58 falls below the 65 convergence bar for new buys, and the specialist explicitly flags the proximity to 52w high as a ceiling on near-term upside. Doesn't quite clear the convergence test. Constructive but not actionable today.

Best price action of the group but win-probability fails the 65 convergence floor for a new buy, and no AMD-specific catalyst is firing. Valuation is extreme and the setup looks like late-cycle chase. Watchlist-only.

Watchlist name fails the new-buy convergence bar (win_prob 58 < 65). Thesis is decent and momentum is the best of the AI chip set in this batch, but proximity to 52w high with stretched valuation caps near-term upside. Not a clear pass on the stock itself, just not a buy here.

Constructive momentum but win_probability=58 fails the 65 convergence bar for new buys, and the 178x PE with proximity to 52w high creates meaningful ceiling risk. Forward expected return is positive but not strong enough to justify initiating against the convergence rule.

Decent momentum and bull narrative but win_probability of 58 fails the 65 convergence bar for new buys. Valuation extreme at 178x PE with stock already near 52wk high limits forward asymmetry. Catalyst is sector-derived not name-specific. Skip — not a convergence setup.

Decent momentum and bull-aligned catalyst but win-probability fails the 65 new-buy bar, and at 178x PE near 52w high the technical ceiling is real. Doesn't meet convergence — skip.

Momentum and catalyst are real, but win-probability of 58 misses the 65 convergence bar for new buys. Valuation is extreme and the stock is already near 52w high with limited room before resistance. Not a clean enough setup to override the rule.

Convergence bar met (thesis 70, win-prob 70) but barely, and the stock is already near resistance with stretched valuation. Bullish setup but not enough edge for a fresh entry given the chip exposure already in the book. Watchlist for breakout confirmation.

Convergence bar requires thesis>=60 AND win_prob>=65 — AMD just clears both at 70/70. But thesis isn't dominant, catalyst is analyst-piece rather than named hyperscaler design win, and stock approaching resistance limits asymmetric upside. Not a clear new_buy when we already own NVDA, AVGO, MRVL in the same lane.

Convergence threshold met (thesis 70, win-prob 70) with a concrete, name-specific catalyst that mirrors the MRVL/MU winner pattern. Per recent learnings, cap the valuation-based risk penalty when a breakout has a real partnership headline behind it. Stock-level call is buy; concentration is Sizer's call.

Decent forward setup but no name-specific catalyst, and competitive signal in Vultr win going to NVDA is a mild negative. Thesis and win-prob both below the 60/65 convergence bar for new_buy. Skip on watchlist.

Decent thesis and momentum but no concrete AMD-specific catalyst — riding generic AI tide near resistance, exactly the pattern our learnings flag for de-rating. Convergence bar for new_buy (thesis>=60 AND win_prob>=65) is just met but quality is weak and catalyst is generic. Pass.

Meets convergence bar (thesis 70, win-prob 70) with concrete named-hyperscaler catalyst (Meta AI chip pact) and Helios rack platform directly attacking Nvidia in AI data centers. Momentum at 96 with near-breakout setup is the pattern that worked on MRVL/MU. Per ARM learning, cap the valuation-based risk penalty when catalyst + momentum + range-position all converge. Sector concentration is Sizer's problem.

Matches the winning pattern: AI-infra hardware with named hyperscaler catalyst (Citi/Meta GPU deal) and strong momentum. Thesis 70 is below the 60 floor but win_probability 70 just meets the bar — convergence is borderline (thesis_pct 70 OK, win_prob 70 over 65). The MRVL learning says cap valuation penalty when catalyst+momentum+breakout align. However thesis_pct of 70 vs other names in batch suggests skip rather than new_buy; conviction reflects the genuine forward setup.

Convergence bar not met — thesis 70 and win-prob 52 both fall short for new_buy. Meta pact is real but valuation leaves no margin and the recent -10.86% week signals exhaustion. Pass.

Meta partnership is real but win probability and risk quality are weak, and thesis at 70 doesn't clear the 60+/65+ convergence bar. Sharp 5d drawdown signals momentum has rolled over. Skip.

Watchlist name fails new-buy convergence rule (thesis 70 ok but win_prob only 52). Sharp 5d breakdown and stretched valuation argue against chasing. Skip.

Fails new-buy convergence bar (need thesis>=60 AND win_prob>=65; win_prob only 52). Extreme valuation and sharp 5d breakdown argue against initiating, especially given existing chip concentration. Skip.

Watchlist name fails the new_buy convergence bar — thesis 70 is at the line but win_prob 52 is well below 65 threshold. Extreme PE, sharp 5d breakdown, and high correlation with existing chip holdings make this a clear skip. No urgency from catalyst justifies action.

Watchlist name failing the convergence requirement — win_probability=52 well below the 65 floor for new buys. Sharp 11% 5d decline, extreme valuation, and weak intrinsic risk score. No reason to chase here.

Watchlist name fails the new_buy convergence test — win_prob 52 is well below the 65 threshold. Thesis is fine but AMD continues to lose AI share to NVDA, and the valuation leaves no room for error. Sector also already concentrated. Skip until momentum stabilizes.

Fails new-buy convergence (win_prob 52 < 65). Sharp 5-day breakdown, extreme valuation, and would worsen sector concentration. No reason to act.

Classic AMD setup from our learnings: thesis_pct=70, catalyst_pct=80, near 52wk high, with a landmark Barclays upgrade directly citing agentic AI. Win-prob at 58 is below the standard 65 floor, but the AMD-specific override applies (thesis+catalyst+momentum+range alignment). Don't let valuation caution veto entry again. Concentration is the sizer's job.

AMD is interesting but fails the convergence bar (win_prob 58 < 65) and momentum_pct=46 doesn't meet the AMD-style breakout override criteria (need momentum_pct >= 75). Constructive watchlist name but no new buy here. Concentration concerns also real.

Per AMD learning, when thesis>=70 AND catalyst>=70 AND momentum>=75 AND range>=95 we override the win-prob floor. Here catalyst is 60 and momentum only 48 — the breakout override conditions are not met. Win-prob 58 falls short of the 65 bar. Constructive watchlist name but doesn't clear convergence today.

Watchlist name with momentum/catalyst breakout profile that matches the AMD-miss learning. However, win_prob=58 and thesis_pct=70 don't clear the convergence bar (60/65). The learning notes accept lower win_prob only when momentum >=75 AND range_position >=95 AND thesis+catalyst >=70 — AMD is borderline but 5d is stalling (-0.38%). Skip for now; revisit on confirmed breakout above $527.

This is the AMD pattern the learnings explicitly call out — thesis>=70, catalyst>=65, momentum>=75, near 52wk high. Prior cycle we let win_probability veto and missed +43%. Applying the documented override: accept win_prob of 58 given the convergence of thesis/catalyst/momentum at a breakout. Risk_quality 25 and sector concentration are real but belong to Portfolio Sizer. New buy with constructive conviction.

Per AMD-learning override, when thesis+catalyst+momentum align at 95%+ range we could relax the win-prob floor — but here momentum_pct is only 70 and the stock is showing -0.38% 5d stall at resistance, not a clean breakout. Without breakout confirmation, the convergence bar isn't quite met. Watch for breakout above $527 as trigger.

Unlike the prior AMD lesson, this read has catalyst going AGAINST AMD (NVDA Arm PC chip explicitly targets AMD's CPU business) and momentum stalling at 52w high. The override conditions don't apply when the catalyst is negative. Skip.

Convergence test fails: thesis 74 clears bar but win_probability 68 just above the 65 floor with risk_quality of 25 and PE 156 near all-time highs. Not a 'great company, great price' setup. Hold off; not adding to chip concentration without stronger conviction.

AMD meets the convergence bar narrowly: thesis_pct=74 and win_probability=68 both above thresholds, with momentum/performance at the top of the batch and a clean breakout pattern that matches our MU/MRVL winning fingerprint. Risk_quality is low on valuation but intrinsic-only; sizer handles concentration. Starter new_buy warranted.

AMD has the momentum and catalyst fingerprint we want (breakout, direct headlines, server TAM raise), and win_prob=68 + thesis=74 nearly clears convergence bar but thesis just misses 60+. However, sizer-level concentration concern is meaningful here given five existing chip holdings. Hold conviction constructively but defer new buy to portfolio sizer.

Strong momentum and direct competitive catalyst, but convergence bar not cleanly met: thesis_pct 74 and win_prob 68 are both below the 'great company, great price' standard for new buys. Risk_quality 25 is weak. Stock has already run 330% off lows — chasing here near ATH without a clean thematic-news cluster headline specific to AMD is exactly the late-cycle entry we want to avoid. Skip.