Pre Market Thoughts - 9 Jul 26

Yields inched higher across the curve. The AI capex build out in my mind is intrinsically tied to US yields as capex driven demand pushes up cost of funding even on the sovereign end. Asia markets today performed weakly relative to the US close.
KOSPI fell as much as 5% from the open to the low, driven by derisking seen in $SKHY as markets start setting the price for $SKHY for friday's NY open. NKY was up 1.4%, a strong showing relative to Korea. It was also revealed that Bain has already disposed of its entire stake in $285A Kioxia. Looking at the chart, it is extremely impressive that such a large order went through, no doubt playing a role in its -40% drop from ATHs at one point.
In HK markets, China semis have made a strong rebound. In my note prior, i talked about Burry's sale of BABA and how that was a bottom signal for $BABA. Well, now we can see how much further BABA can rally. Remember, BABA is integrated into almost every aspect of China's AI push.
In US trading yesterday, markets staged a strong rebound as I highlighted that semi fund flows to make space for the $SKHY subscription was finishing up. We saw this in particular retail favourite names like $NBIS, which started leading the neoclouds again in performance. Notably, the sell twap in $NBIS seems over for now.
Away from flows, $META also unveiled the building of a new 1GW datacenter in Canada. This proves my gut instinct on the $META headline headfake last week - the hyper scalers are going to ramp up capex.
From a positioning angle however, i am less bullish on memory. It seems to me that the memory trade is very well understood now and very heavily positioned already by market participants. My initial thesis for neoclouds and memory pooling to lead was based off companies finding work arounds into memory and also from a positioning perspective. This has been reaffirmed.
On Neoclouds, we have 2 catalyst that dropped that i believe will result in what i term as neocloud summer. Firstly, $NVDA's new credit support scheme to select neoclouds is going to help the market by allowing for a cheap cost of funding. It also breaks the stranglehold that the hyperscalers have on the neoclouds. This is because alot of neocloud financing on the debt side now, favors hyperscalers because they can lend their credit to the neoclouds - this allows the hyperscalers to pay a cheaper rate for compute. $NVDA in this case has broken that stranglehold that the hyperscalers have now. Neoclouds can not only lower their cost of funding BUT also be able to secure higher paying clients.
Secondly, Anthropic turning profitable INCLUDING TRAINING in 3Q destroys the bear case that all of this neocloud, frontier lab financing is a ponzi scheme. Anthropic has now turned into a cash flow generating monster with north of 70% gross margins on serving inference. Currently, they are limited on compute and this will just tighten the entire compute market harder.
On markets - i also repeatedly warn people to manage their risk well. Only when you do so, would you be able to take advantage of market dips to add to positions vs being forced out of a position. Semis while being the structural game in town also faces overly long positioning - this creates conditions for flushes and gaps in liquidity.
On the geopolitical front, the US-Iran spate looks likely to continue for a few weeks. CL is lower today than from yesterday, i think that being long cheap tail call spreads can provide fantastic protection for portfolios in the event that this erupts further.
Good luck!