Original Views Summary

🧠 AI Memory & Storage

• Memory may be a great long-term setup, but the key overhang is not economics. It is politics. The hard question: why should storage capture half of the total investment from the biggest industrial revolution in human history as profit? Until that political question has an answer, buying the sector is still hard to underwrite. 1

• $DRAM and $DISK are not the same trade. $DRAM is the cleaner bet on the memory cycle itself: HBM shortage squeezes DRAM capacity, which then tightens DDR5 / server supply. $DISK is more like an “AI storage supply-chain enhancer,” with heavier exposure to NAND, SSD, HDD, and Asian storage suppliers. 2

₿ Crypto Market Structure

• $BTC absorbed a clear negative catalyst. $MSTR announced it sold $200 million of Bitcoin, and BTC only stabbed lower briefly before snapping back. That de-sensitization matters. The prior setup was that if BTC survived the fragile final two days of late June, a rebound could follow. That read is now playing out. 3

🤖 AI Software, Cybersecurity & Biotech

• The market may be misreading software because $IGV hides major dispersion. The stronger pocket is cybersecurity and AI-agent permission management: $PANW, $CRWD, $FTNT, $NET, $OKTA. $PANW and $CRWD have been strong enough to rival peak $DRAM-style momentum. Meanwhile, heavyweights like $MSFT and $ORCL distort the aggregate software tape. 4

• The Fable 5 framework effectively restricts model usage in two areas: cybersecurity and biotech. That makes those two sectors structurally different from generic AI beneficiaries. The cleaner read is to revisit the stocks in those verticals, because the model-access constraint can become part of the moat. 5

• If the market is crowded with CAPEX anxiety, the cleaner rotation may be into $AAPL and $PLTR rather than staying stuck in the most CAPEX-sensitive names. 6

🏭 Semiconductors & WFE

• The semi tape is at a real make-or-break level. $SOXX and $SMH are testing the 50 day moving average / 50dma for the first time since the massive rally began in early April. Bulls likely need to defend soon, or the trend loses momentum. 7

• UBS’s SPCX Terafab call is a huge potential WFE signal: $50 billion of WFE spend in 2030/2031 alone. That is basically like adding another TSMC. If that capex path is real, the trade is to buy WFE now, before the spend is fully priced. 8

• Hynix is too large in market cap to be played like a random small-cap 13F pump. The usual pattern—build a position first, trigger retail FOMO through 13F visibility, then distribute into the crowd—is much harder to execute at that scale. 9

🏦 Fintech & Brokerages

• If the new account flow also goes through $HOOD, it could become another major stock tailwind. The market has already seen how powerful distribution narratives can be through the Trump children accounts angle. 10

• $SOFI is trading at a fresh all-time low relative to $HOOD. That does not mean $HOOD must sell off. The cleaner relative-value read is that $SOFI is increasingly due for outperformance. Target: back to the mid $20’s by eoy or sooner. 11