OpenAI’s own figure to investors was approaching $50 billion at the end of September, counted its way: Azure resale booked net, only OpenAI’s cut. Investors then added the estimated cloud-partner revenue so the total could be lined up with Anthropic, which books the full customer bill through AWS and Google and treats the platform fee as a cost. That add was about $20 billion, which is how the late-September reports got to roughly $70 billion.

The S&P 500 and Nasdaq are trading less than 1% from their all-time highs. Although the major indexes largely avoided the typical August–September correction, significant weakness occurred beneath the surface. Many individual stocks corrected below key moving averages, breadth deteriorated, and equal-weighted indexes weakened while mega-cap strength supported the capitalization-weighted indexes. In effect, the market experienced a stealth correction, stock by stock. Evidence of that stealth correction can be seen clearly in market breadth. Despite the major indexes trading at or near all-time highs, the percentage of stocks trading above their 50- and 200-day moving averages has fallen to historically low levels. In fact, we have seen some of the lowest participation above these key moving averages ever recorded with the major indexes at or near all-time highs. That is an unusually large divergence between what the headline indexes are showing and what is actually happening beneath the surface. That could provide an interesting setup entering a historically favorable Q4, particularly during a midterm election year. If breadth improves and previously corrected stocks stabilize and regain their key moving averages, participation could broaden and strengthen the advance. Even without significant broadening, historical seasonality still favors the major indexes moving higher into year-end. Small-cap stocks, emerging markets, and many cyclical areas have lagged, leaving market gains increasingly concentrated among a handful of large-cap leaders. The resilience of AI-related themes, despite broader market weakness, continues to support the case for technology-oriented exposure. The primary complication remains interest rates and inflation. Nominal GDP growth remains strong, but historically the 10-year Treasury yield has averaged roughly 200–250 basis points above the GDP deflator; recently, that spread has been closer to 100 basis points. Either inflation needs to moderate and validate current yields, or Treasury yields may still have further to rise. A 10-year yield peaking near 5.25%, particularly alongside moderating inflation, would be constructive for equities. A decisive move above that level with inflation remaining elevated could create a significant headwind. Investors are also becoming increasingly discriminating about credit risk, with signs of widening spreads between lower-quality high-yield debt and investment-grade bonds. While the move is not yet indicative of a full-blown credit stress event, it represents another potential warning developing beneath the surface. When investors begin demanding greater compensation for owning lower-quality debt while equities remain near record highs, it can signal that risk perceptions are changing before those concerns become evident in the major stock indexes. For now, the trend remains positive, but selectivity is important. Focus on high-quality stocks emerging from sound bases with strong or improving relative strength and favorable price-volume characteristics. Most important, cut losses on all stock that fail, and only get aggressive on the heels of traction. NO FORCED TRADES. NO BIG LOSSES. The key question for Q4 is whether improving breadth and favorable seasonality can reinforce the uptrend—or whether persistent inflation, rising yields, and renewed distribution undermine it.