Signal UNCHANGED — 🚫 DO NOT BUY across all three horizons. Price near-flat (+0.6% ($386.09 → $388.57)), but the name is more expensive on the anchor, not less: the 10-Y rose 4.48% → 4.67%, lifting the discount rate to 9.2% and trimming the warranted multiple 27x → ~26x, so clean 64.8x is now 2.46x warranted (was 2.36x). The 'S&P 500 concentration / AI earnings-quality unwind' tail was re-confirmed ARMED in the 30 Jul macro (top-10 ≈ 41% of the index; RSP flat while XLK ripped +5.5% on 29–30 Jul — breadth narrowing), so DNB Trigger 2(b) still fires. Economic Alignment: medium-term pressure flips Tailwind → Headwind (XLK medium Underperform in the 30 Jul macro); long stays Tailwind (XLK Outperform). Quality flat 86, driver flat 78. Timing nudged 44 → 47 (MTF confluence flipped bearish → bullish; price reclaimed the 200-DMA). Nothing in the business deteriorated — this remains a price-and-risk call.
Broadcom is a global technology company built from two engines. Roughly two-thirds of it is semiconductors — most importantly the custom AI accelerators (XPUs/ASICs) it co-designs for hyperscale cloud operators and the Tomahawk/Jericho networking silicon that wires AI data-centre clusters together, alongside wireless, broadband and storage chips. The other third is infrastructure software, anchored by VMware (data-centre virtualisation) plus mainframe and cybersecurity franchises. Its edge is scale, deep customer integration and relentless cost discipline under CEO Hock Tan: it buys critical-infrastructure franchises, raises their margins and turns them into durable, high-cash-flow annuities. For a reader: think of Broadcom as one of the few 'picks-and-shovels' giants selling into both the AI hardware build-out and the enterprise software stack at once.
Lifecycle: Growth (accelerating). Broadcom is a hybrid — roughly two-thirds semiconductors (AI custom-silicon/XPU for hyperscalers, AI networking, wireless, storage) and one-third infrastructure software (VMware, mainframe, security) — so it is scored on a blended lens: Semiconductor gross-margin + utilisation on the chip side, software recurring-revenue / Rule-of-40 economics on the VMware side. Q2 FY26 (quarter ended 3 May 2026) revenue was $22.19B, +47.9% YoY — a genuine acceleration driven by the AI franchise, not a mix trick. TTM revenue ~$75.5B, TTM operating income $32.9B, TTM net income $29.3B.
| Sub-signal | Value | Read | Score |
|---|---|---|---|
| Revenue trajectory | +47.9% YoY (Q2 FY26) | Accelerating on AI custom-silicon + full VMware run-rate | 92 |
| Gross margin (TTM) | 67.0% | Strong for a fabless/software hybrid; software mix lifts it | 88 |
| Operating margin (TTM) | 43.7% | Elite operating leverage; Hock-Tan cost discipline | 90 |
| Net margin (TTM) | 38.8% | High, and CLEAN (non-op is a drag, not a boost) | 86 |
| FCF (TTM) | ~$27.2B, FCF/sh $6.90 | Cash conversion >90% of op cash flow | 85 |
| Balance sheet | Net debt ~$45B; interest cover 11.0x; current 2.24 | Levered from VMware but comfortably serviced; de-levering on plan | 70 |
| ROE (TTM) | 33.4% | Top-decile; ROIC ~88th percentile of large-cap semis | 88 |
Moat score = 73. Switching-cost and cost-advantage sub-scores are trimmed directly from the competitive read below — stable share against credible rivals, not runaway dominance.
| Rival | Threat type | Share trajectory | Moat-erosion vector |
|---|---|---|---|
| Nvidia (NVDA) | Merchant GPU vs custom ASIC — the substitute at the platform level | Stable — different buying motion; hyperscalers run BOTH | If merchant GPUs get cheap/abundant enough, the case for bespoke ASIC narrows at the margin |
| Marvell (MRVL) | Direct custom-silicon / ASIC competitor for the same hyperscaler programs | Stable to slightly ceding at the top end — AVGO holds the marquee accounts | The only credible second source for a custom XPU program — caps AVGO pricing power |
| Arista (ANET) / Cisco (CSCO) | AI-networking / switching — the systems layer above the silicon | Stable — AVGO is the merchant silicon INSIDE many of their boxes (Tomahawk/Jericho) | Hyperscaler in-house switch ASICs are the longer-run nibble |
| Hyperscaler in-house teams (Google TPU, AWS Trainium, Meta MTIA) | Vertical substitution — the customer becomes the designer | Stable — AVGO co-designs many of these, so it captures the work rather than losing it | The structural tail risk: a hyperscaler internalises the full stack and drops the merchant partner |
ROIC & capital allocation: ROIC in the top decile of large-cap semis (~88th percentile), ROE 33.4%. Capital-allocation score 80 — disciplined M&A (VMware being de-levered and margin-expanded on plan), a growing dividend (payout ~40%), buybacks. Management skin-in-the-game 60 — Hock Tan's alignment is strong, but SBC and the sheer size of past deals warrant a neutral-plus, not a maximal, score.
Warranted-multiple anchor — the decisive lens. A company is the present value of its future cash flows: growth (g) lifts the warranted multiple, the discount rate (r) — which rises with interest rates — lowers it.
| Input | Value | Note |
|---|---|---|
| Risk-free (10-Y UST) | 4.67% | From FRED DGS10, 29 Jul 2026 — up from 4.48% at the prior report |
| Equity risk premium | 4.50% | Fixed global constant |
| Risk add-on | +0.0% | Business-Quality 86 ≥ 65 → no add-on |
| Discount rate r | 9.2% | 4.67 + 4.50 + 0.0 |
| g_near (yrs 1–5) | 14% | Consensus growth haircut 25%, held near the 15% secular-growth cap despite +48% prints |
| g_term (yr 6+) | 3% | Long-run nominal GDP cap |
| Warranted P/E | ~26.3x | Two-stage DCF; below the 28x semis guardrail |
| Actual clean P/E | 64.8x | Price $388.57 / TTM diluted EPS $6.00 |
| Ratio (actual ÷ warranted) | 2.46x | ≥ 1.40 → EXPENSIVE; deeper than the prior 2.36x |
The forward-P/E "mitigant" is exactly the hype the anchor is built to discount. Street models EPS roughly tripling by FY27 (some provider estimates put FY27 revenue near $175B vs ~$75B TTM — not credible at face value), which is how you get a ~20x "forward P/E." The anchor haircuts that optimism by design; on clean trailing earnings the name is 2.46x warranted. Do not let a forward multiple built on tripled earnings launder a 64.8x trailing multiple into "cheap."
Earnings quality — CLEAN, and if anything the multiple is understated. Non-operating items are a net drag (interest expense -$745M in Q2 FY26; total other income/expense -$739M) — there are no mark-to-market gains on private-AI stakes inflating EPS. TTM net income $29.3B sits below TTM operating income $32.9B, so clean_pe = reported P/E (nonop 0%). The TTM effective tax rate of 3.8% (a Q4-FY25 one-off benefit) actually flatters trailing EPS; normalised to ~14% tax, the trailing P/E would be ~70x+. No downward normalisation is warranted — the multiple is clean and deep-expensive.
| Cross-check | Reading |
|---|---|
| Analyst consensus target | $506.5 (median $520, high $582, low $400) — +30% to consensus; 51 Buy / 8 Hold / 0 Sell (86% bullish, 59 covering) |
| EV/EBITDA (TTM) | 45.0x — rich vs semi peers even adjusting for software mix |
| P/S (TTM) | 24.4x |
| FCF yield | ~1.5% — thin for the multiple |
The relative lenses can only order the name within the Expensive band the anchor sets — they cannot lift it out. Valuation score 25.
Primary driver: AI custom-silicon (XPU/ASIC for hyperscalers) + AI networking, with the VMware infrastructure-software annuity as the secondary. This is one of the two or three cleanest large-cap expressions of the AI-capex build-out: Broadcom co-designs bespoke accelerators for multiple hyperscalers and supplies the Tomahawk/Jericho switching silicon that stitches AI clusters together.
| Horizon | Score | Read |
|---|---|---|
| Historical | 90 | AI revenue has compounded hard; custom-ASIC bookings ramped through FY25–FY26 |
| Current | 76 | Demand intact, order book strong; but the trade is crowded and consensus already prices a supercycle |
| Forward | 72 | Multi-year hyperscaler capex visible, yet capex-digestion / in-house-silicon risk caps the forward read |
Driver score 78 — Strong Tailwind. Amplification is moot here: the base signal is HOLD (High Quality + Expensive) and the final signal is DO NOT BUY (Trigger 2b), so a strong tailwind never reaches the base to lift it — a driver can intensify a BUY, never rescue a gated/DNB name. Thesis-invalidation floor: a sustained hyperscaler AI-capex cut or custom-silicon order push-out that breaks the AI-revenue acceleration narrative.
Technology → XLK carries a split macro read in the 30 Jul MacroDriver report: short Neutral · medium Underperform · long Outperform. So the pressure on Broadcom is Neutral in the short run, a genuine Headwind medium-term (down from Tailwind at the prior report — the AI-concentration tail is armed and breadth is narrowing), and a Tailwind long-term (the secular AI build-out). Buying an Expensive AI-cohort mega-cap into a medium-term sector Underperform + an armed concentration tail is a Contrarian stance, hence conviction is a middling 60.
Source: sector-map · Macro report 2026-07-30
The tape is constructive on the longer timeframes and choppy intraday — but timing is a secondary consideration when the name is gated Expensive and DNB. Price $388.57 sits above the 200-DMA ($365.78) but below the 50-DMA ($395.74); RSI(daily) 50.9 is neutral, daily MACD -2.34 (histogram just turned up +0.45).
| Sub-signal | Value | Read |
|---|---|---|
| MTF confluence | Bullish (longer-TF) | Monthly + weekly + hourly uptrend; daily/15-min weakening — flipped from bearish at the prior report |
| Risk-reward | 48 | Reward to consensus is real, but entry above both the June low and the anchor's fair band is poor |
| Relative strength | In line | Recovered off the June low (~$360); roughly in line with XLK (short Neutral) |
| Catalyst cluster | 68 | Q3 FY26 earnings 3 Sep 2026 is the next real catalyst (>14d out) |
| Dynamic macro weight | 0.15 | Semiconductors = Medium macro-sensitivity |
Timing score 47 (up from 44). A better tape does not change the call: an Expensive name with an armed systemic de-rating catalyst is a DO NOT BUY regardless of momentum.
| Date | Event | Impact | Forecast | Previous | Relevant? | Why |
|---|---|---|---|---|---|---|
| 2026-08-12 | US CPI (Jul) | High | — | — | Indirect | Rate path feeds the discount rate on a long-duration semi multiple |
| 2026-09-03 | Broadcom Q3 FY26 earnings | High | EPS ~$3.22 / rev ~$29B (est) | Q2 $1.91 dil | Direct | AI-revenue trajectory + custom-ASIC bookings — the key catalyst; >14d out |
| 2026-09-16 | FOMC decision | Medium | — | — | Indirect | Semis = Medium macro-sensitivity; rate path feeds the anchor |
| Date | Event | Actual | Forecast | Surprise | Impact |
|---|---|---|---|---|---|
| 2026-07-29 | 10-Y Treasury yield | 4.67% | — | +19bp vs prior report | Lifts r to 9.2% → warranted multiple 27x → ~26x → more Expensive |
| 2026-07-30 | Macro AI-concentration tail | armed | — | re-confirmed | Keeps DNB Trigger 2(b) live; top-10 ≈ 41% of index, breadth narrowing |
The only stock-specific catalyst of note is Q3 FY26 earnings on 3 Sep — beyond the 14-day window, so the next scheduled refresh is the default +14d (14 Aug). The macro backdrop (armed concentration tail, 10-Y at 4.67%) is what pins the signal.
| Timeframe | Trend | Direction | RSI | MACD | Key S/R | Breakout | Vol |
|---|---|---|---|---|---|---|---|
| Monthly | Uptrend | ↑ | 63.7 | +2.8 hist | R 414.6 | resistance_breakout | 0.8x |
| Weekly | Uptrend | → | 52.7 | −4.4 hist | S 321 / R 414–495 | resistance_breakout | 0.79x |
| Daily | Weakening | → | 50.9 | −2.34 | S 360 / R 414 · 50-DMA 395.7 | below 50-DMA | 1.13x |
| Hourly | Uptrend | ↑ | 54.3 | +1.83 | R 399.9 | resistance_breakout | — |
| 15-min | Weakening | ↓ | 50.0 | −0.06 | S 379.7 | support_breakdown | — |
| Confluence: Bullish (longer-TF) · mixed intraday · MTF Score 58 | |||||||
Longer timeframes are constructive (monthly/weekly/hourly uptrend, price back above the 200-DMA at 365.8) while the daily and 15-min are weakening under the 50-DMA at 395.7. Net: the tape has healed since the June low but sits in a $360–$414 range. Immaterial to the call — the name is gated Expensive and DNB regardless of the chart.
Indicative 6-month daily close (synthesised from known anchors: 52-wk low $281.61, a $414 high, the June low ~$360, 50-DMA $395.7, 200-DMA $365.8, spot $388.57). For visual context only.
The AI-capex supercycle sustains and broadens: custom-ASIC wins accelerate across multiple hyperscalers, AI networking share holds, VMware margins keep expanding. The market keeps paying a premium multiple as earnings compound into it. Price runs toward the Street high (~$582). ~+44%.
The franchise compounds but the multiple grinds lower as the 10-Y stays elevated and the sector digests capex. EPS growth (toward ~$19–20 on FY27 Street) offsets modest de-rating; price drifts up toward consensus-minus. ~+17%. Most probable.
The COHORT MULTIPLE-DE-RATING leg (mandatory, and armed): the macro AI-concentration / earnings-quality tail fires — a hyperscaler capex guide-down or an AI private-markdown event — and the whole AI-capex cohort re-rates. AVGO's 64.8x compresses toward the low-20s on FY26 earnings; a Marvell/in-house-silicon share scare adds a company-specific leg. ~−45%. This is a live, not hypothetical, left tail — which is precisely why the signal is DO NOT BUY rather than HOLD.
Probability-weighted fair value ≈ $421 (0.25·560 + 0.50·455 + 0.25·215), which sits above today's $388.57. That is deliberate and not a contradiction of the DO-NOT-BUY call: DNB is a risk-override, not a modal-target call. Two things sit on top of the positive modal path — (1) the Valuation-Ceiling hard gate caps the name at HOLD on price alone, and (2) Trigger 2(b) overrides to DO NOT BUY because the −45% bear is an armed systemic catalyst (the 30 Jul macro's AI-concentration tail), not a hypothetical. When a fat, live left tail can erase two-plus years of upside in one re-rating, the risk-adjusted entry is poor even though the mean is positive. We do not chase a name that is obviously overpriced AND carrying a live puncture risk.
Forecast: No entry path opens while the name is Expensive AND the AI-concentration tail is armed. The realistic route back to even a HOLD is a de-rating: price into the ~$300s (toward the anchor's fair band) OR the macro report disarming the concentration tail (breadth broadening). Neither is imminent. Re-tested at the 14 Aug refresh.
Forecast: Not applicable — this is a DO-NOT-BUY, not a held position. The exit ladder is shown for completeness; no trigger is live.
No fresh position is warranted at this price (DO NOT BUY), so no sizing is computed. Position sizing guidance would require your portfolio allocation and role — and in any case does not apply to a gated/DNB name.
{
"ticker": "AVGO",
"exchange": "NASDAQ",
"exchange_ticker": "NASDAQ:AVGO",
"isin": "US11135F1012",
"api_ticker": "AVGO",
"finder_ticker": "AVGO",
"finder_exchange": "\ud83c\uddfa\ud83c\uddf8 NASDAQ",
"section": "Technology (Contrarian)",
"country_table": "US",
"date": "2026-07-31",
"version": "v6",
"analysis_status": "on-going",
"user_context": {
"horizon": null,
"allocation_pct": null,
"portfolio_role": null
},
"user_horizon": null,
"user_allocation_pct": null,
"portfolio_role": null,
"company": "Broadcom Inc.",
"currency": "USD",
"price_at_rating": 388.57,
"lifecycle_stage": "growth",
"sector": "Technology \u2014 Semiconductors + Infrastructure Software (hybrid)",
"signal_short": "DO NOT BUY",
"signal_medium": "DO NOT BUY",
"signal_long": "DO NOT BUY",
"primary_signal": "DO NOT BUY",
"composite_short": 46,
"composite_medium": 54,
"composite_long": 62,
"quality_score": 86,
"quality_confidence": 80,
"quality_detail": {
"industry_benchmark_name": "Semis GM+Utilisation (67% GM) blended with Software recurring-revenue/Rule-of-40",
"industry_benchmark_value": 67,
"industry_benchmark_score": 88,
"moat_score": 73,
"roic_percentile_vs_peers": 88,
"capital_allocation": 80,
"management_skin_in_game": 60,
"gross_margin_ttm": 0.6696,
"net_margin_ttm": 0.3885,
"ebitda_margin_ttm": 0.5569,
"operating_margin_ttm": 0.4366,
"roe_ttm": 0.334,
"revenue_growth_latest_q_yoy": 0.479
},
"valuation_score": 25,
"valuation_confidence": 85,
"val_band": "expensive",
"val_multiple_basis": "clean P/E",
"actual_multiple": 64.8,
"warranted_multiple": 26.3,
"warranted_ratio": 2.46,
"discount_rate_r": 9.2,
"risk_free_10y": 4.67,
"g_near": 14,
"g_term": 3,
"sector_guardrail_pe": 28,
"clean_pe": 64.8,
"clean_peg": 0.91,
"actual_pe_ttm": 64.6,
"valuation_detail": {
"pe_ttm": 64.6,
"forward_pe_fy27": 19.9,
"peg_ttm": 0.5,
"forward_peg": 0.91,
"ev_ebitda_ttm": 45.0,
"ev_revenue_ttm": 25.5,
"price_to_sales_ttm": 24.4,
"price_to_book_ttm": 21.0,
"fcf_yield": 1.5,
"historical_valuation_decile": 7,
"boundary_note": "EXPENSIVE on the warranted-multiple anchor \u2014 clean P/E 64.8x (TTM diluted EPS $6.00, price $388.57) vs a warranted ~26.3x (two-stage DCF: r 9.2% = 10-Y 4.67% + 4.5% ERP + 0.0% add-on for BQ>=65, g 14% near / 3% terminal), ratio 2.46x, AND above the 28x Semiconductors guardrail floor. Either condition fires EXPENSIVE (score < 40); both fire. The 10-Y rose 4.48%->4.67% since the prior report, trimming warranted 27x->26x, so on a near-flat price the name is MORE expensive (ratio 2.36x->2.46x). Even at max g_near=20% the guardrail still caps it. High Quality + Expensive -> HOLD base; STRONG-BUY amplification blocked."
},
"nonop_pct_of_net_income": 0,
"earnings_quality_note": "CLEAN \u2014 non-operating items are a net DRAG (interest expense -$745M in Q2 FY26; totalOtherIncomeExpensesNet -$739M), no mark-to-market equity-stake gains. TTM net income $29.3B sits BELOW TTM operating income $32.9B, so reported EPS is not inflated; clean_pe = reported P/E. TTM effective tax 3.8% (Q4-FY25 one-off benefit) flatters trailing EPS; on a normalized ~14% tax the trailing P/E would be ~70x+ (richer). No downward normalization warranted.",
"timing_score": 47,
"timing_confidence": 65,
"timing_detail": {
"mtf_confluence_score": 58,
"mtf_tool_confluence": "bullish",
"trend_monthly": "uptrend",
"trend_weekly": "uptrend",
"trend_daily": "weakening",
"trend_hourly": "uptrend",
"trend_15min": "weakening (support_breakdown)",
"risk_reward_score": 48,
"relative_strength_vs_spy": "recovered off the June low (~$360) to ~$389; roughly in line 1m/3m",
"relative_strength_vs_sector": "in line with XLK (short N)",
"rsi_daily": 50.9,
"macd_daily": -2.34,
"sma50_daily": 395.74,
"sma200_daily": 365.78,
"atr_daily": 16.29,
"catalyst_clustering_score": 68,
"dynamic_macro_weight": 0.15,
"macro_sensitivity": "Medium (Semiconductors)"
},
"driver_score": 78,
"driver_label": "Strong Tailwind",
"driver_confidence": 65,
"driver_detail": {
"primary_driver": "AI custom-silicon (XPU/ASIC for hyperscalers) + AI networking; secondary: VMware infrastructure-software annuity",
"historical": 90,
"current": 76,
"forward": 72,
"amplification_eligible": true,
"amplification_applied": false,
"amplification_blocked_reason": "Base signal is HOLD (High Quality + Expensive) and the final signal is DO NOT BUY (DNB Trigger 2b), so the driver tailwind is moot \u2014 amplification never reaches HOLD, and DNB overrides everything.",
"thesis_invalidation_floor": "A sustained hyperscaler AI-capex cut / custom-silicon order push-out that breaks the AI-revenue acceleration narrative"
},
"economic_alignment_stance": "Contrarian",
"economic_alignment_conviction": 60,
"economic_alignment_pressure": "Neutral",
"economic_alignment_pressure_short": "Neutral",
"economic_alignment_pressure_medium": "Headwind",
"economic_alignment_pressure_long": "Tailwind",
"economic_alignment_source": "sector-map",
"economic_alignment_confidence": 66,
"macro_report_date": "2026-07-30",
"macro_sector_signal": "XLK s:N m:U l:O",
"competitive_share_trajectory": "stable",
"competitive_threat_level": "moderate",
"moat_score": 73,
"overall_confidence": 65,
"fair_value_est": 315.0,
"stop_loss": 353.0,
"support_levels": [
365.78,
360.46,
321.42,
289.96
],
"resistance_levels": [
395.74,
414.64,
442.36,
495.0
],
"target_price": 506.5,
"analyst_consensus_target": 506.5,
"analyst_target_high": 582,
"analyst_target_low": 400,
"analyst_target_median": 520,
"analyst_target_upside_pct": 30.4,
"analyst_grades_consensus": "Buy",
"analyst_bullish_pct": 86.4,
"analyst_coverage_count": 59,
"recent_upgrades_30d": 0,
"recent_downgrades_30d": 0,
"fmp_rating": "B",
"fmp_overall_score": 3,
"hard_gate_state": "triggered",
"gates_triggered": [
"Gate 3 \u2014 Valuation Ceiling: clean 64.8x is 2.46x its ~26x warranted multiple AND above the 28x semis guardrail line. Either arm fires; both fire. Caps High Quality + Expensive at HOLD and blocks STRONG-BUY amplification."
],
"gates_caution": [],
"do_not_buy_triggers": [
"Trigger 2(b) \u2014 Valuation Extreme (absolute arm): the name is DEEP in the Anchor's Expensive band (clean 64.8x = 2.46x warranted, above the 28x semis guardrail) AND a material, currently-ARMED systemic de-rating catalyst is live \u2014 the macro report's 'S&P 500 concentration / AI earnings-quality unwind' tail (status: armed, stamp 2026-07-30). AVGO materially belongs to that cohort: its multiple and revenue acceleration are directly levered to the hyperscaler AI-capex / custom-silicon trade. Arm (b) has NO 'exceptional proven growth' exemption. Earnings are CLEAN (nonop 0%), so the earnings-quality half of the tail does not apply, but the AI-capex-leverage half does \u2014 one arm of the cohort test is sufficient. Expensive + a live index-level de-rating catalyst = obviously overpriced AND risky -> DO NOT BUY, all three horizons."
],
"dnb_trigger_2b_fired": true,
"entry_groups_met": 0,
"entry_conviction": "Wait",
"exit_groups_live": 0,
"exit_action": "Hold",
"short_entry_confirmed": false,
"short_hold_reason": "expensive",
"short_cap_reason": "DNB Trigger 2b override (Expensive + armed AI-concentration tail)",
"scenario_bull_target": 560,
"scenario_base_target": 455,
"scenario_bear_target": 215,
"scenario_probs": {
"bull": 25,
"base": 50,
"bear": 25
},
"next_update_date": "2026-08-14",
"next_check_date": "2026-08-14",
"next_update_basis": "default +14d (no dated catalyst before Q3 FY26 earnings 2026-09-03, which is >14d out; Semiconductors = Medium macro-sensitivity, so the 16 Sep FOMC is not a scheduling trigger). The armed AI-concentration tail is a standing DNB condition, re-tested each refresh.",
"last_updated_human": "Jul 31, 2026",
"prior_report": {
"date": "2026-07-16",
"price_at_rating": 386.09,
"signal_short": "DO NOT BUY",
"signal_medium": "DO NOT BUY",
"signal_long": "DO NOT BUY",
"quality_score": 86,
"valuation_score": 26,
"timing_score": 44,
"driver_score": 78,
"economic_alignment_pressure": "Neutral",
"economic_alignment_conviction": 60,
"warranted_ratio": 2.36,
"hard_gate_state": "triggered",
"do_not_buy_triggers": [
"Trigger 2(b) \u2014 Valuation Extreme (Expensive + armed AI-concentration tail)"
]
}
}
Every score, level and the full gate/DNB state is captured in the JSON above so the next run (14 Aug) can compute deltas. The load-bearing fields: val_band=expensive, warranted_ratio=2.46, hard_gate_state=triggered, dnb_trigger_2b_fired=true.