NEUTRAL / CAUTIOUS Friday's broad bounce lowered immediate stress, but did not repair a losing week or the renewed rise in long Treasury yields.
Generated 08:34:12 CEST (Europe/Rome) · Data cutoff 08:34 CEST / 06:34 UTC. U.S., European and Asian cash figures are Friday closes; U.S. futures are not treated as actionable while regular weekend trading is closed. This Saturday edition audits prior calls without hindsight and maps the next week.
All four U.S. equity benchmarks rose and VIX fell, but Treasury's official 2Y and 10Y each climbed 5 bp to 4.24% and 4.74%; the 30Y rose to 5.27%. That is tactical relief, not a clean risk-on confirmation. Friday recap · Treasury
NVDA fell 0.98% to $214.72 while S&P and Nasdaq rose. Broadcom gained 1.21%. The divergence raises the hurdle for Wednesday's NVIDIA earnings to validate demand, margin and financing assumptions. NVDA quote · AVGO quote
RGTI +11.48%, QBTS +8.46% and IONQ +8.02% sharply outperformed. No new primary-company economic disclosure was verified alongside the move, so treat it as high-beta/speculative breadth rather than improved unit economics. RGTI · QBTS · IONQ
| 2026-08-21 call | Pre-stated confirmation / invalidation | Friday evidence | Paper outcome |
|---|---|---|---|
| NVDA conditional rebound long | Reclaim $219–220 with breadth and 10Y below 4.67%; invalid below $215.66. | High $218.74, close $214.72; 10Y 4.74%. | Not triggered Long condition never occurred. Price broke invalidation, so setup is retired—not counted as an activated loss. |
| AMZN conditional bearish | Below $259.50 with weak Nasdaq; invalid on reclaim $264. | Low $257.04, close $258.63, but Nasdaq +0.44%. | Not triggered Price leg occurred without required weak-index confirmation. No hindsight short credit. |
| AMLX avoid / bearish only below offering | No bearish entry unless below $35.50; invalid above $40.44. | Range $38.35–$41.14; close $38.66. | No entry Avoid stance preserved capital; neither bearish trigger nor durable invalidation confirmed. |
| MU relative-strength watch | Hold $950 and clear $1,000 on volume; invalid below $930. | Low $958.26, high $989.96, close $966.78. | Pending Support held, breakout did not occur. |
| AVGO watch | Fresh quote plus close above $368; invalid below $355. | Fresh close $368.45, +1.21%; high $375.13. | Confirmed Trigger cleared by $0.45. This becomes the only constructive carried setup; earnings risk remains. |
| XBI / QBTS / WMT watches | XBI close above $166; QBTS fresh price/volume but no numeric entry; WMT several sessions above $103 then $108 reclaim. | XBI $165.73; QBTS +8.46%; WMT $103.70. | No formal trigger QBTS moved sharply, but yesterday defined observation—not a trade—and no retrospective entry is assigned. |
Risk class: large-cap, high-duration AI/custom silicon · Classification: confirmed watch-to-long candidate · Horizons: day trade only with Monday follow-through; 1–4 week swing into Sep. 2; 3–12+ month only after earnings. Direction remains uncertain across the event.
Thesis/catalyst: Friday closed $368.45 (+1.21%), above the pre-stated $368 confirmation, while NVDA/MU fell. Sep. 2 earnings can validate custom silicon and networking demand.
Evidence / counterargument: Relative strength and a confirmed close support the setup. Counter: intraday rejection from $375.13, 10Y at 4.74%, customer concentration, integration/leverage and earnings-gap risk.
Entry/confirmation: hold $365–368 Monday and reclaim $372; avoid chasing a gap. Invalidation/downside: close below $355; next risk near $345. Key risks: estimates, AI capex, margins, rates, customer concentration.
Risk class: mega-cap, high-duration / binary earnings gap · Classification: avoid new pre-earnings swing; tactical bearish only on confirmation · Horizons: day trade Monday–Wednesday; 1–4 weeks after results; structural 3–12+ month thesis undecided.
Thesis/catalyst: $214.72 Friday, −0.98% against Nasdaq +0.44%, and below yesterday's $215.66 invalidation. Wednesday's report is too binary for an unconfirmed long.
Evidence / counterargument: Weak relative strength and rising yields argue for caution. Counter: profitable scale, documented contracted demand and a strong report can gap through resistance.
Entry/confirmation: bearish only below $214.50 with weak semis; do not hold a tactical short through earnings. Invalidation/downside: reclaim $219–220; bearish downside references $210 then $200. Short loss is unlimited; no borrow claim.
Risk class: small/high-beta speculative growth; dilution and squeeze prone · Classification: avoid chase, not an active basket short · Horizons: day trade only; 1–4 week observation; no 3–12+ month entry without commercial evidence. Direction uncertain.
Thesis/catalyst: +8% to +11% Friday without newly verified economic disclosure creates poor chase asymmetry under a 4.74% 10Y.
Evidence / counterargument: Absolute volume and breadth were strong. Counter: momentum can persist violently, and commercial deployments may create legitimate option value.
Entry/confirmation: no short entry Monday; only consider bearish exposure after a failed Friday-high retest and break of Friday low (RGTI $16.30, QBTS $19.10, IONQ $42.01). Invalidation: closes above Friday highs. Risks: squeezes, dilution, architecture shifts, contract headlines; short losses unlimited.
No second compelling bullish idea: Friday breadth improved, but yields rose and the week's central macro/earnings catalysts create unfavorable gap risk.
| Ticker / risk class | Classification & separately applicable horizon | Thesis, catalyst & evidence / counterargument | Entry / invalidation / downside / risks | C / R / O |
|---|---|---|---|---|
| Watch MU Nasdaq · liquid cyclical AI memory | Rebuild watch. Day: above $990. Swing 1–4 weeks: only above $1,000. Long term: wait for cycle/valuation update. Direction uncertain. | Held $950 but failed $1,000 Friday; NVIDIA can validate memory demand. Counter: −0.77% lag and memory cyclicality. | Confirm close >$1,000 with volume; invalidate <$950, harder stop <$930. Supply, pricing, capex and geopolitical risks. | 67 / 79 / 62 |
| Watch XBI NYSE Arca · diversified high-volatility biotech ETF | Stabilization. Day: $168 break. Swing: close >$168. Long term: diversified sleeve only. Direction uncertain. | +1.44% to $165.73 and less binary than AMLX. Counter: failed to close above prior $166 trigger and funding costs remain high. | Confirm >$167.91 Friday high; invalidate <$163.74 then $160. Clinical, FDA, pricing and financing risks. | 65 / 70 / 60 |
| Watch AMLX Nasdaq · small/mid-cap clinical biotech; binary/dilution sensitive | Post-offering base watch. Day/swing only after base; 3–12 months speculative after NDA path. Direction uncertain. | Phase 3 endpoint/statistics and reported safety support avexitide; $500m gross offering improves runway. Counter: 14.09m new shares, approval/label/manufacturing/launch unknown. | Confirm close >$41.14 then $42; invalidate <$38.35, major downside reference $35.50 offering. Binary gaps, competition and dilution. Not penny/OTC. | 82 / 91 / 59 |
| Watch AMZN Nasdaq · mega-cap capex-intensive cloud/platform | Repair watch. Day: $261 reclaim. Swing: close >$264. Long term: only with FCF/capex visibility. Direction uncertain. | AWS growth is a structural counterweight; Friday low held $257. Counter: −0.57% on an up tape and high AI capex sensitivity. | Confirm $261 then $264; invalidate <$257.04, downside $250. Rates, capex, competition, consumer and regulation. | 66 / 72 / 58 |
| Watch WMT NYSE · mega-cap defensive retail | Post-earnings stabilization. Day: hold $103. Swing: reclaim $108. Long term: reassess margins/guidance. | Friday closed $103.70 after Thursday's 9.15% fall, with low $102.15. Counter: one flat session is not a base. | Confirm multiple closes >$103 then $108; invalidate <$102.15/$100. Margin, consumer, tariff and guidance risks. | 61 / 64 / 56 |
Bias: neutral until Monday breadth and bonds agree. AVGO is constructive only above $372 after holding $365–368; avoid chasing quantum gaps.
Biggest risk: weekend geopolitical/tariff gap. Invalidation: broad index weakness with 10Y above 4.75% shifts bias risk-off; 10Y below 4.67% plus positive breadth shifts risk-on.
Bias: keep size light into Wednesday. Prefer AVGO confirmation; wait for post-earnings evidence in NVDA and for XBI/MU breakouts.
Biggest risk: GDP/PCE and NVIDIA jointly reprice rates and the AI complex. Invalidation: S&P close above 7,700 with 10Y below 4.67% would improve the cautious stance; below 7,660 with rising yields worsens it.
Bias: profitable AI infrastructure remains preferable to quantum pure plays or single-asset biotech, but new capital should wait for cash-flow/valuation evidence. Use speculative sleeves only for quantum/biotech.
Biggest risk: paying peak expectations as discount rates rise. Invalidation: durable FCF acceleration, lower long yields, or regulatory/clinical success that materially improves probability-adjusted cash flows.