Pre Market Thoughts - 17 Jul 26

Brutal morning in asia trading. Korea markets are closed today. NKY was down as much as 6% at one point in early morning trading, a sign that the panic contagion from NY markets has indeed spread into Asia. I look at this via the terminal ( where you can customize the tickers you add in and this makes the relationships between price moves in ticker to really pop.
$285A is down 16% today to 52K JPY. I see strong support for this at about 49-52k zone. In overnight markets, memory tickers like $SNDK have reached 1320 and $MU near 800s. These are at my buy zones and i have deployed my first tranche of remaining cash here.
Today is also OPEX day, so expect funky price action going into NY open.
How am i thinking of trading this market now - first, this is the most vicious dip in semis this year. If you (like me) believe that the AI supercycle is not over, then this is probably a good time to deploy cash. Next, i am cognizant that the sell off could get worse if there is a further rout on the index level as the semi sell off sparks a general derisking move. This is why i recommended to add some index puts yesterday which were trading at very cheap levels. Personally i did a 7.5/7.35/7.2 P/fly for next week's expiry.
In terms of interesting flows, there was a big clip of SKHY $200 calls that were done through the market yesterday.
On fundamentals. Kimi has released K3 - a model that on front end engineering tests scored just behind GPT and Claude. This is the first time that the Chinese labs have pushed out a model that is on par with the US. Diving into the details, we find that the cost per task for K3 is actually the same as GPT.... This hints to me that alot of the training for K3 probably relied on distillation, otherwise they should also show similar gains in lowering the inference costs. The market in today's trading has decidedly weighed that this is bearish for AI. However, if Kimi releases these open weights, i would argue that it is bullish for the Capex buildout, because this will drive inference demand. This might not be good for OAI and Anthropic but i am also certain that the US frontier labs have much stronger models already.
Today is also the last day of a trading week before big tech earnings (check the calender , toggle earnings at if you want to see the dates)
Fortune favors the bold.
Good luck