It’s been a strange weekend in the AI world. Anthropic CEO Dario Amodei called for a slowdown in AI development, and Elon Musk agreed with him. The CEO of OpenAI has made similar comments lately, and we’ve now seen two Anthropic employees quit over concerns about AI safety. Meanwhile, President Trump and Palantir CEO Alex Karp argued that slowing down would just hand China an advantage in the AI race.
If you take the weekend’s comments at face value, either AI is going to take over the world, or China is. If that is not stuck between a rock and a hard place I do not know what is.
Stocks showed continued weakness last week, and really for the past month. Given the seasonality we’re up against, the breaks below key averages in GOOGL and AMZN, and the fact that QQQ hasn’t made a new all time high since June 4th and just can’t push higher, I’m growing increasingly cautious.
If QQQ opens below the 100 day SMA, that’s a meaningfully bearish gap, and I’d expect a fill of the $701 gap this week. That’s a solid level of support, and we could see a bounce there. But after failing at the 100 day SMA with this much uncertainty weighing on the market, I expect the pullback to continue over the next month.
QQQ Daily Chart
The market typically doesn’t perform well at the start of a rate hiking cycle, which likely begins Wednesday. There have been five rate hiking cycles in the SPYs existence and the SPY always saw at least a 7% drawdown within the first two months. The good news: those pullbacks have historically created some phenomenal buying opportunities.
2022: SPY pulled back 19% over 211 days from the first hike to the market low.
2015: SPY pulled back 13% over 35 days from the first hike to the market low.
2004: SPY pulled back 7% over 38 days from the first hike to the market low.
1999: SPY pulled back 10% over 111 days from the first hike to the market low.
1994: SPY pulled back 8% over 18 days from the first hike to the market low.
Every one of these drops turned into a buying opportunity for traders willing to step in, and several were phenomenal opportunities even on a longer-term horizon. The one exception from a long-term perspective was 1999. The market topped just 158 days after the low we discussed above, followed by the Dot-com crash, which took the SPY down 50% over the next 931 days. However, even in that case, the SPY rallied another 26% off the hike low before eventually topping and crashing.
SPY Daily Chart
If history repeats and we get a September and October drop, which seasonality certainly points toward, that pullback should be buyable heading into year end, even if the AI bubble is about to burst.