美国股票策略：AI风险溢价的演变

Hyperscalers, neo-clouds, and data center SPVs have been funding the historic AI DC buildout through sources like IG & HY debt, convertibles, and equity. We measure the evolution of the risk priced in these assets, through credit spreads, CDS, equity risk premium, and equity vol metrics. Equity risk premia on hyperscalers have been volatile throughout this year, but have sharply declined from the year-to-date peak over the last two months. This has occurred despite strong earnings beats in 2Q26, upward revisions in NTM estimates, and de-rating valuations. The key driver has been the surging long term yields, with the 10-year US treasury yield rising 70 bps in this period to 5.28%. Long term implied volatility on hyperscalers appears rich in relation to realized volatility and equity index volatility, with historically flat skew and relatively low put-to-call open interest and volume ratios hinting at continued demand for upside. Over the past two quarters, with unprecedented credit sup