Rewired Index · Market · Today

Today

截至 2026-09-24

数据暂未更新，以下是上一份完整数据

今日读数

稳定观察

经济数据未见转弱，VIX 与参与广度未落在极端档。

判断 2026-09-24 · 部分数据已过期
初请失业金四周均线
202.2K强
10 年期美债收益率
5.162%承压
收益率曲线 2s10s / 3m10y
+26bp走平
VIX 水平与 20 日方向
15.67常态
S&P 500 成分股高于 50 日均线比例
27.0%受损
QQQ 与长期趋势线
741.10趋势向上
AI 资本开支健康度
$165.1B数据不足
前沿模型能力推进
+16.6 ECI / 365D仍在推进
美国企业 AI 使用率
22.4%使用面扩大

什么情况下这个判断不成立

任一条件出现即推翻当前判定：初请失业金不再同时满足「四周均线与最新值都低于 260.0K」和「四周均线不上升」，但也没有连续 4 周上升（现 四周均线 202.2K）→ 橙灯；初请失业金四周均线连续 4 周上升（现 四周均线 202.2K）→ 橙灯；初请失业金四周均线升破 300.0K 且连续 4 周上升（现 四周均线 202.2K）→ 红灯；VIX 升破 28 且 20 日涨幅达到 25%，或 S&P 500 高于 50 日均线的比例跌破 25%（现 VIX 15.67 · 广度 27.0%）→ 绿灯；四大云厂合计资本开支代理环比低于 -15%，或 NVIDIA 数据中心收入连续 2 个季度下降（现 四大云厂代理 +27.2% QoQ）→ 情景 C

下一个该看的数据点

2026-09-30 · PCE

宏观

观测分歧 观测 2026-09-24

就业与利率分歧

指标走势读数对照状态观测日
初请失业金四周均线
<260K 且平稳或回落 强 · 连续四周上升 转弱 · >300K 且上升 恶化
202.2K最新 197K强2026-09-19
10 年期美债收益率
<4% 或缓慢下行 支持 · 4–4.5% 中性 · >4.5% 或月升 40bp 承压
5.162%20 日 +52.3bp承压2026-09-24
收益率曲线 2s10s / 3m10y
≤0bp 倒挂 · 0–50bp 走平 · >50bp 陡峭
+26bp20 日 -21bp走平2026-09-23

分歧 · 10 年期美债收益率提示压力，但初请失业金四周均线、收益率曲线 2s10s / 3m10y尚未同向确认

风险偏好

观测分歧 观测 2026-09-24

风险偏好分化

指标走势读数对照状态观测日
VIX 水平与 20 日方向
<20 常态 · 20–28 恐惧 · >28 且上行 极端
15.6720 日 +3.0%常态2026-09-24
S&P 500 成分股高于 50 日均线比例
>60% 健康 · 40–60% 中性 · <40% 受损 · <25% 崩塌
–27.0%50D 27.0% (136/503) · 20D 25.6% (129/503) · 200D 47.5% (238/501)受损2026-09-24
QQQ 与长期趋势线
高于 50D 与 200D 趋势 · 低于 50D 高于 200D 修正 · 低于 200D 熊市风险
741.1050D 711.20 · 200D 663.53趋势向上2026-09-24

分歧 · S&P 500 成分股高于 50 日均线比例提示压力，但VIX 水平与 20 日方向、QQQ 与长期趋势线尚未同向确认

AI 健康

数据已过期 观测 2026-09-09 · 含过期数据

观测已经过期

指标走势读数对照状态观测日
AI 资本开支健康度
四大云厂合计 QoQ < -15% 或 NVIDIA 数据中心收入连续两季下降
–$165.1B四大云厂代理 +27.2% QoQ数据不足2026-06-30
前沿模型能力推进
365 天推进 ≥ 8 点 仍在推进 · < 3 点 基本停滞 · 之间为放缓
–+16.6 ECI / 365D150.0 → 166.6 ECI仍在推进2026-09-09
美国企业 AI 使用率
13 个双周期变化 ≥ +1.0pp 扩大 · ≤ -1.0pp 收缩 · 之间为持平
–22.4%13 个双周期 +4.9pp使用面扩大2026-08-09

缺数据 · 资本开支数据不全；缺推理成本数据；企业采用数据已过期；缺能源约束数据

Rewired Index · Market · Today

Today

As of 2026-09-24

Not updated yet; showing the last complete data

Today's read

Steady · keep watching

Economic data shows no weakening, and neither VIX nor breadth sits in an extreme band.

Read 2026-09-24 · some data is stale
Initial claims, 4-week average
202.2KStrong
10-year Treasury yield
5.162%Pressure
Yield curve 2s10s / 3m10y
+26bpFlat
VIX level and 20-day direction
15.67Normal
S&P 500 members above the 50-day average
27.0%Damaged
QQQ vs its long-term trend
741.10Trending up
AI capex health
$165.1BInsufficient data
Frontier model capability progress
+16.6 ECI / 365DStill advancing
US business AI adoption
22.4%Expanding

What would overturn this

Any one of these overturns the current verdict: initial claims stop meeting both "four-week average and latest print below 260.0K" and "four-week average not rising", without yet rising for 4 consecutive weeks (now four-week average 202.2K) -> Orange; the initial-claims four-week average rises for 4 consecutive weeks (now four-week average 202.2K) -> Orange; the initial-claims four-week average breaks above 300.0K while rising for 4 consecutive weeks (now four-week average 202.2K) -> Red; VIX breaks above 28 with a 20-day gain of 25%, or the share of S&P 500 members above their 50-day average falls below 25% (now VIX 15.67 · breadth 27.0%) -> Green; the four-hyperscaler capex proxy falls more than 15% quarter over quarter, or NVIDIA data-centre revenue declines for 2 consecutive quarters (now hyperscaler proxy +27.2% QoQ) -> Scenario C

Next data point

2026-09-30 · PCE

Macro

Observations diverge Observed 2026-09-24

Jobs and rates diverge

Indicator (framework thresholds)Trend · main series, thin line = comparisonReadingComparisonStateObserved
Initial claims, 4-week average
<260K and flat or falling: strong · rising four weeks in a row: weakening · >300K and rising: deteriorating
202.2KLatest 197KStrong2026-09-19
10-year Treasury yield
<4% or drifting lower: supportive · 4–4.5%: neutral · >4.5% or +40bp in a month: pressure
5.162%20d +52.3bpPressure2026-09-24
Yield curve 2s10s / 3m10y
≤0bp: inverted · 0–50bp: flat · >50bp: steep
+26bp20d -21bpFlat2026-09-23

Divergence · The 10-year Treasury yield signals pressure, but initial claims (4-week average) and the 2s10s / 3m10y curve have not confirmed in the same direction

Risk appetite

Observations diverge Observed 2026-09-24

Risk appetite is split

Indicator (framework thresholds)Trend · main series, thin line = comparisonReadingComparisonStateObserved
VIX level and 20-day direction
<20: normal · 20–28: fear · >28 and rising: extreme
15.6720d +3.0%Normal2026-09-24
S&P 500 members above the 50-day average
>60%: healthy · 40–60%: neutral · <40%: damaged · <25%: collapse
–27.0%50D 27.0% (136/503) · 20D 25.6% (129/503) · 200D 47.5% (238/501)Damaged2026-09-24
QQQ vs its long-term trend
Above 50D and 200D: trending · below 50D, above 200D: correction · below 200D: bear-market risk
741.1050D 711.20 · 200D 663.53Trending up2026-09-24

Divergence · The share of S&P 500 members above the 50-day average signals pressure, but VIX (level and 20-day direction) and QQQ vs its long-term trend have not confirmed in the same direction

AI health

Data is stale Observed 2026-09-09 · includes stale data

Observation is stale

Indicator (framework thresholds)Trend · main series, thin line = comparisonReadingComparisonStateObserved
AI capex health
Big-four hyperscaler capex QoQ < -15%, or NVIDIA data-centre revenue down two quarters in a row
–$165.1BHyperscaler proxy +27.2% QoQInsufficient data2026-06-30
Frontier model capability progress
≥8 points over 365 days: still advancing · <3 points: stalled · in between: slowing
–+16.6 ECI / 365D150.0 → 166.6 ECIStill advancing2026-09-09
US business AI adoption
Change over 13 fortnights ≥ +1.0pp: expanding · ≤ -1.0pp: contracting · in between: flat
–22.4%13 fortnights +4.9ppExpanding2026-08-09

Missing · capex data incomplete; no inference cost data; enterprise adoption data is out of date; no energy constraints data