One interesting situation our Blue Horseshoe tool flagged early was $GRNY buying a big chunk of $MSTR in mid-Aug which caused it to go from the last stock in the port to the top holding in two wks flat (prior to a 46% run in the stock so good call). Also notable bc of the GRNY's ether connection.

One of the more instructive trades of my career was one that should have worked and did not, and it taught me what a put spread actually is. February 2018. I had a put spread on the book. The market sold off four or five percent in a violent move, the sort of day protection is bought for. I looked at my position expecting a decent number and the put spread had barely changed in value. My honest first reaction was to wonder how that was even possible. A hedge, in a proper sell-off, doing nothing. The answer is that a put spread is a short skew position. You are long a put nearer the money and short one further out. Those lower strikes are where the short vega is concentrated in a move down, and when vol and skew explode the way they did that week, the option you are short gains more than the one you are long. Skew steepened violently against me and it swallowed the gain the delta was handing me. There is a second problem stacked on top. A put spread only pays its maximum when the short strike decays to nothing, which means at expiry. Sell-offs happen fast and bounce fast, so the moment your protection theoretically looks best is usually a moment you cannot cash it at. If the market had settled at the lows, the skew would have come back in and I would have got the money. Markets do not tend to do that. So put spreads are not a cheaper version of a put. They are a different position, with a different exposure, that behaves worst in exactly the conditions you bought it for. I'm not saying don't buy them, I'm just saying know what you're buying when you do.

One thing I have kept coming back to with $AVGO is that I thought the AI TAM could end up much larger than people were modeling, because Broadcom was never really just a custom XPU story. The same customers building their own accelerators also need the networking around them. Switch silicon, SerDes, optical DSPs, PCIe, lasers. And as the clusters get larger, Broadcom gets more content around each dollar of compute. That is why I enjoy reading the report. Broadcom now expects AI semiconductor revenue to reach ~$115B in FY27 and ~$230B in FY28, while saying AI networking should grow roughly as fast as XPUs over the next few years. We used to discuss the possibility that the TAM was being underestimated. I think we are starting to see that happen in the actual numbers. Google is getting bigger. Anthropic is becoming huge. OpenAI has moved from roadmap to shipments. Meta keeps progressing. And Broadcom gets paid on more than one layer of all of it. That is pretty much the exact reason $AVGO has remained one of my highest-conviction holdings