Broad Uptrend, Quiet Adapt-Divergence-Reading AI InfrastructureNVIDIA CorporationBATS:NVDATrendMatrix_LabsNVDA closed at 222.82 today, holding above its smoothed signal path with the Adaptive Trend Filter reading **Good** across every structural metric — **Strong** momentum at +1.38% over the last 5 bars, **Increasing** trend pressure, and **Perfect** Technical Rating. The MAMA/FAMA cycle filter is tracking an 18-bar dominant cycle with the Reversal Engine quiet (no conviction score above 50). The broader screener reports the NASDAQ AI Infrastructure universe in **Broad Uptrend** — 23 of 37 tickers in Elite tier, only 2 in Bear. The headline read is unambiguous. The interesting story is what sits underneath it. ### The Elite cohort is dense and stacked in semis The top 10 Elite names by rank are exclusively chips, semi equipment, and the SMH benchmark itself: - **#1 SNDK** (41b in tier) — RankVal +84.2, ADD · Strengthening - **#2 MU** (35b) — +39.8, ADD · Strengthening - **#3 DELL** (27b) — +33.1, ADD · Strengthening - **#4 MRVL** (32b) — +32.5, ADD · Strengthening - **#5 ARM** (15b) — +26.1, ADD · Strengthening - **#6 AMAT** (53b) — +21.2, HOLD · Mature - **#7 LRCX** (38b) — +20.5, HOLD · Stable - **#8 AMD** (13b) — +19.4, ADD · Strengthening - **#9 KLAC** (51b) — +15.7, HOLD · Mature - **#10 SMH** (40b) — +13.4, ADD · Strengthening Eight of ten read **ADD · Strengthening**; two read **HOLD · Mature** without rank decay. That's a textbook breadth pattern — the secondary semis are running with concentrated conviction, and the long-standing trends (AMAT at 53 bars, KLAC at 51 bars) are holding without fatigue. Notable: **NVDA itself ranks below #10 in the Elite tier**. The namesake of the AI trade is in Elite, but it's the secondary semis — memory, equipment, custom silicon — driving the rank-momentum right now. That's a regime-internal observation worth holding in mind. ### The Confirmed tier is flashing one specific warning Eight names sit at 2-of-3 bullish — and every one shares the *same* failure mode. The Adaptive filter (MAMA/FAMA) has flipped bearish while FAST and CORE remain bullish: - **NYSE:VRT** — 6b, A↓-7, WATCH · Adapt divergence - **NYSE:GEV** — 9b, A↓-3.9, **TRIM · Adapt warn+fading** - **NYSE:PWR** — 1b, A↓-0.6, WATCH · Adapt divergence - **NASDAQ:EQIX** — 11b, A↓-0.5, WATCH · Adapt divergence - **NYSE:DLR** — 12b, A↓-1.6, WATCH · Adapt divergence - **NYSE:TT** — 10b, A↓-1.0, WATCH · Adapt divergence - **NASDAQ:AMZN** — 1b, A↓-0.7, WATCH · Adapt divergence - **NYSE:HUBB** — **20b, A↓-1.5, WATCH · Adapt divergence** ← Top Read The screener's "Top Read" line surfaces HUBB specifically: 20 bars of Confirmed-tier persistence with MAMA flipped bearish. That's the longest-running adapt-divergence in the cohort and the engine's flagship warning. ### Why this matters The Adaptive filter detects cycle changes through MESA spectral analysis — it tends to flip first when the dominant market cycle is shifting, before the slower Hybrid (CORE, 3-pole) and SuperSmoother (FAST, 2-pole) filters confirm. When the *same* failure mode appears across multiple Confirmed-tier names — A↓ while F↑ and Co↑ — that's the regime engine telling you something specific: the second-tier of the AI infrastructure trade (electrical infra, cooling, data center REITs, power gen) is starting to lose cycle support, and the slower filters haven't caught up yet. This doesn't invalidate NVDA. The chart subject is in Elite tier with Perfect Technical Rating, Strong momentum, and a 2.73% Safety Margin above its baseline. But the screener's value here is making visible what raw price action obscures — that "23 in Elite, 2 in Bear" headline doesn't capture eight Confirmed names simultaneously flashing the same cycle warning. ### The Emerging tier confirms the pattern from below Three names read **WATCH · Persistent early** with Adapt-only-bullish while FAST and CORE remain bearish: - **NASDAQ:SMCI** — 35b, A↑+20.4 (rank trending up) - **NASDAQ:MSFT** — 34b, A↑+4.3 (rank trending up) - **NYSE:ORCL** — 35b, A↑+12.1 (rank trending up) These are names where the cycle filter has been calling improvement for 34-35 bars before the structural trend confirms. Patient setups, not noise. ### What invalidates this read NVDA breaking its CORE baseline (visible on chart as the green-to-red transition zone) would flip its consensus to Confirmed and put the Adaptive filter's divergence behavior on the namesake — a different question than where it currently sits. More directly: if HUBB's adapt-divergence persists to 30+ bars *and* its rank trajectory turns negative, the engine would re-tag from WATCH to TRIM · Adapt warn+fading, and the Confirmed-tier warning shifts from observational to actionable. Conversely, if MAMA flips back bullish across the Confirmed cohort and the Adaptive filter rejoins consensus, the warning resolves and the Broad Uptrend regime extends without the qualifying footnote. ### Bottom line Multi-filter consensus screening isn't about replacing trend judgment — it's about making *internal* sector state visible at a glance. Right now, the AI infrastructure stack has strong headline breadth and a synchronized cycle warning sitting one tier beneath it. Both are real. Both matter for how you size and where you trim. --- *Indicator: Ehlers Adaptive Trend Filter + Consensus Engine by TrendMatrix Labs* *Not financial advice. Position sizing, entry timing, and risk management are independent of trend identification.*