What it measures
The Capital Deployment Posture Index reads how the public corpus is allocating dollars across the cycle. When AI capex and growth-investment language dominates buyback language, the index rises — companies are deploying. When buyback language dominates, the index falls — companies are returning capital, often a signal that they don't see better internal uses for it. The number is multiplied by the public-sentiment component so that a high deployment posture during low confidence is dampened (real growth investment requires confidence to mean anything).
Why these components
AI mentions today is the single largest growth-capex signal in S&P 500 discourse — every major operating company is making AI investment statements. Buybacks are the cleanest counter-signal — when a company is buying back its own stock at scale, it is signaling that it cannot redeploy capital at a better risk-adjusted return. The ratio of the two captures the deploy-vs-return decision the corpus is making at any given moment. Sentiment-weighting prevents the index from reading a low-confidence pivot to growth talk as a true growth-mode read.
How to read it
Above 70 — strong deployment posture; the corpus is choosing growth. Between 40 and 70 — balanced; normal capex-to-buyback ratios. Below 40 — defensive posture; the corpus is returning capital. The index is most actionable as a regime indicator paired with credit spreads — when posture is deteriorating and spreads are widening, capital is being returned defensively rather than deployed.
What it does not say
The index doesn't distinguish between productive and unproductive AI capex. Companies can over-invest in growth and still register high on this index. It's a regime indicator, not a quality indicator. Pair with capex-to-revenue ratios on the underlying components to separate the two reads.
History note
History recomputed 2026-09-21 on the rebuilt five-year Discourse corpus under the canonical discourse.* definition: readings before 2026-09-19 are weekly as-of recomputes that replace the history originally computed on the pre-rebuild corpus.