Market backdrop
The recent market swings come straight from earnings reactions, and many AI-related names are pulling back. I read that pullback as setup, not breakdown. We're heading into what should be a strong fall and winter for the stock market and the AI sector.
The Nvidia chief joined X today and posted the exact thesis I've been pushing since January and repeated last week: open AI models are coming. These open models will put heavy pressure on OpenAI, Anthropic, and others. A lot of people are misreading this and getting scared, which is dragging AI names down. I see it the opposite way. The pressure forces more innovation, and my rule holds: never short American. If OpenAI's and Anthropic's leaders can't take the heat, they can leave the kitchen.
AI hardware is the place to watch. When the Kimi model was released, it had to be shut down the same night because there wasn't enough hardware to run it. That is the demand signal. Avoid the fear and look for the good ideas.
Alphabet (GOOGL)
I called Google as my top AI idea for 2026 back in September or October, and it worked. It has now pulled back to support after a mixed, ugly earnings reaction, and it's slightly higher today. If it bounces here and figures things out, which I trust it to do, this pullback to recent resistance turned support could be a great entry. The warning: you're only as good as your last hit. Yahoo was once the best thing going, and Google buried it. Google has to deliver now or risk the same fate.
The chart is weaker in the short run. Former support near 340 has flipped to new resistance. Price keeps resolving lower since earnings, sits well below its 50-period moving average on the hourly chart, and is in a bearish momentum regime despite some short-term steadiness at these lows. It looks like a possible bear pennant is forming. If that consolidation triangle completes, the measured move points another $20 lower in the very short term.
Breaking market news
Mid-segment, the market turned. The Russell went green, up a third of a percent; the NASDAQ was the only major index in the red, down four-tenths. The move followed unconfirmed reports that Pakistan is considering returning to stalled US-Iran talks to end a nearly five-month war, reportedly after a push from China. Exploratory talks took place during Iran's interior minister's visit to Islamabad this week, his second in the last 10 days. Markets would welcome real progress.
Super Micro (SMCI)
This was one of the early AI picks years ago, well before its big run. It gapped down hard on odd headlines about smuggling chips, then filled that gap right back to pre-news levels. Last time it filled a gap, it ran well from there. The demand case is simple: hyperscalers can't buy enough hardware, and Super Micro is the household name in data center plays. Barring more strange headlines, a repeat of the move from a couple weeks ago looks doable. The stock is up about 25% over the last week on a $60 billion backlog it announced.
The technicals show a clear split across time frames, so this is not the highest-probability setup. Short term, most of the period was a downtrend until a large decisive candle on Tuesday broke resistance and flipped it to a bullish regime. The MACD still sits on the zero line, with price just above old resistance now acting as new support around $30. A higher high just formed, and a higher low usually follows, so a pullback to the midpoint of that big Tuesday candle around 27 would confirm a possible structural uptrend or reversal.
On the daily chart, momentum is still bearish, though it has carved a higher high and higher low. The main wall is the anchored VWAP from former pivot highs near 32; clearing that puts bulls in control toward 50 or 51. The key level for bulls is holding Wednesday's gap, the window from 26 to 2850. If that window closes within a few bars, price could slide back to the yearly lows. Over the last 52 weeks the stock is still down 42%. It's currently near 3026, down 3% on the day.
CVS
CVS is my flight-to-safety pick, a balance to the more speculative AI ideas for people not ready to touch a Google or an SMCI right now. It's a "boring" name that's up about 100% in a year, just trotting along in a steady trend. At this point it has done no wrong, so you stay long. When people get scared, they move to safer places like healthcare and banks.
The trend is your friend until it bends. Short term, CVS has stair-stepped higher in a strong structural uptrend, with old resistance becoming new support at 106. If 106 fails, the next levels down are 102, then 96. Over the last year it was rangebound most of the time, then broke out into a relentless uptrend. On the daily chart it just printed a tweezers top just above the 261.8% Fibonacci retracement, a typical profit-taking zone, and it's trading right at that level. Momentum has been fading for a couple months, shown by a large bearish MACD divergence, which is a yellow warning or profit-taking sign. The trend hasn't broken yet. If price can't hold the Fibonacci level near 107, longer-term investors will watch the 200% Fibonacci around 98, which lines up with the short-term support. CVS is up close to 1% at 10782.
