Either way, it should bode well for semicaps especially with recent $TSM results. Asian readthroughs also positive

So built out a semicap long basket over the past week:

$ONTO @ $$221
$BESI @ $$172
$ASM @ $700
$MKSI @ $212
$NVMI @ $462

Out of $INTC from $43.8 to $51 - ugly print no two ways about it. While they’re likely sandbagging their guide/commentary, mgmt just can’t sell a narrative. Stock also carries a Trump association risk, which I think increases going into midterms. Probably better to express the INTC/foundry view through semicaps, and you can even get higher beta despite diversification.

Net ~55% gain on the name, not too shabby but probably too high an oppy cost. I’m never married to a stock, if the thesis changes I’m out.

Have been building a long position in $PSTG around the $70 ish range. One of the most overlooked stocks for the memory trade imo- I misunderstood this name as well previously as a short but profited through luck.

Market probably perceives it as similar to commodity vendors like Dell/Netapp. But I think their DirectFlash IP will become extremely invaluable in a tight NAND supply scenario in increasing density & resilience through a software-first approach. And I think NAND demand is about to blow through the roof.

$MU just announced NAND capacity expansion in their SG fab today, slated to come online in 2028. Market is interestingly, rewarding the NAND players in light of the news, instead of punishing them. You can guess why

My memory picks are $SNDK (trimmed into print, will buy after print), $285A.JT, $PSTG

Started long positions in Agentic AI software, $NET $DOCN $DDOG $GTLB

Anyone who has spent time playing with Claude Code/Clawdbot this month comes to 2 conclusions: 1) Agentic (and useful) AI is finally here, and will likely accelerate from here into the mass market realm. 2) there is and will be infinite demand for compute.

Software has been beaten down so hard and so long, some names more undeservingly so than the others. And now there’s a high rate of change secular tailwind thats about to dawn on these names, all while positioning heavily leans towards the short side. As clean of a narrative as it gets

I was actually very interested in $NET at ~$200 last year. But valuation then was extremely high, and I concluded that the trade was too early and thus exited. Fast forward a few months, valuation has came down but it is still very high.

The thesis has however gotten more pristine and surprisingly simple to understand: Agentic AI is going to increase web traffic by 10x+, as agents relentlessly trudge along while you sleep. Now imagine Cloudflare as a tollbooth situated right in-front of a large % of global sites. Thats all.

The stock can grow into its valuation imo, but a key risk for me is that the founder is kinda.. not very keen on maximising monetisation potential. I still like the stock and setup is clean, but whether I hold it for long is ultimately dependent on whether the mgmt changes their mindset.

$STX CEO from print earlier

“Our near-line capacity is fully allocated through calendar year 2026, and we expect to begin accepting orders for the first half of calendar year 2027.”

“Multiple cloud customers are discussing their demand growth projections for calendar 28. Underscoring that supply assurance remains their highest priority.”

“Agentic AI relies on persistent access to large volumes of historic data to enable effective planning, reasoning and independent decision making.”

“We believe the stage is set for a sustained and meaningful increase in data generated and stored that will support inferencing, continuous training, and also maintain model integrity.”

Pushing up other memory/storage players too

- inference era is here, while we aren’t even done with training (Blackwell models aren’t even out yet)

- memory is the new bottleneck

Added more $285A.JT @ $19330

From Seagate results, it’s evident that the market is starting to look past CY26 and is focusing more on CY27++. Hence Kioxia shouldn’t be overly punished for its 2026 LTA with Apple, but rewarded for its higher exposure to eSSD in CY27+

Besides I have always liked the wabi sabi ones more

$ASML bookings almost 2x estimates.. If you've been paying attention, a beat isnt a surprise. The extent is though. It's Semicap's party.

Short attack on $CVNA

https://x.com/gothamresearch/status/2016547974776861083?s=46

Got me thinking, what if the recent precious metal rally in Gold/Silver wasnt just pricing in loose monetary policy/geo tensions, but also (maybe more so) pricing in this scenario where infinite printable money ceases to have any more meaning and value in a world of infinite productivity?

Elon has touched on this point a fair bit, maybe he isnt trolling. In that case, hard assets (PMs, land) would accrue all the value.. idk, I have no horse left in the race - just letting my mind ponder, and maybe a trade idea comes out of it all

Q2 Results:
Revenue: $3.025B (Est. $2.67B)
EPS: $6.20 (Est. $3.49)

Q3 Guide:
Adj. EPS: $12.00–$14.00 (Est. $4.21)
Non-GAAP Gross Margin: 65.0%–67.0%

$SNDK just guided Q3’26 EPS to be higher than the FULL FY26 EPS consensus estimates from just 3 months back… stock +16% AMC

https://x.com/wallstengine/status/2016983139344929046?s=46

Given this trend, $SNDK FY27 EPS can reach >$80; up more than 5x from $15 consensus estimates just a few months ago.. power of exponential pricing and 100% passthrough to margins.

The counterpoint is demand destruction from consumer side, but the counterpoint to that is hyperscalers are price insensitive and will gobble anything remaining up, especially with NVDA ICMSP.

We’ll see, but if a $80 FY27 EPS is possible then FY+1 P/E is.. 8x. The stock got cheaper as it rallied.