Braden Brock

Investors have spent much of the past week debating the revenue numbers coming out of OpenAI and Anthropic, trying to determine what they mean for the AI trade. I wonder whether we’re paying too much attention to those numbers and not enough to what’s happening in the bond market. U.S. 10-year Treasury yields have risen for seven consecutive months. In data going back to 1962, we’ve seen a streak this long only once before, in the late 1970s. We went back through history to examine what happened after extended periods of rising yields. In eleven comparable episodes, yields generally paused or declined for periods ranging from two to twelve months. In four cases, the streak actually marked the peak in yields. I find that interesting because of what happened in the bond market recently. Following the latest inflation report, the 10-year briefly reached 5.30percent before reversing to close around 5.18percent. Despite another inflation surprise, the bond market struggled to sustain higher yields. I’ve argued for higher yields for some time, and I haven’t abandoned that longer-term view. Strong nominal growth, persistent fiscal deficits, and enormous demand for capital could keep upward pressure on interest rates. But after seven consecutive months of increases, the possibility of a pause deserves attention. And that could matter considerably more for the AI trade than the latest revenue forecasts. Semiconductor stocks have experienced four separate corrections of 10 percent or more since peaking in June. Forward valuation multiples have contracted roughly 35 percent, even as earnings estimates have continued rising. The market has become considerably less willing to pay for that growth, despite continued strength in demand for compute. The broader equity market has experienced something similar. The S&P 500’s forward valuation multiple has declined from roughly 23 times to 19 times, while earnings estimates have continued rising. If bond yields stabilize or begin declining, investors may become more willing to pay higher multiples for those earnings. Positioning could reinforce that move. Hedge fund net exposures have fallen to their lowest levels since the Liberation Day selloff, despite the S&P 500 trading near record highs. If yields begin declining and equities move higher, investors who have reduced exposure may find themselves chasing the market again. For now, I’m watching whether the recent reversal in the 10-year yield marks the beginning of a period of consolidation. We could see yields decline 40–50 basis points without changing the longer-term argument for higher interest rates. But even a temporary pause could be enough to change the direction of equities. That strikes me as a more useful signal for the next leg of the AI trade than another debate about how quickly the labs are growing their revenue.