EdmundK

OCBC down 10% in two days after Citi sell call with DBS and UOB sliding is classic capitulation, not fundamentals. US 10Y touching 5.35% is peak yields, banks actually earn more NIM here. This washout on 10-share lot launch week is a gift entry for 6% yield.

Record S&P/Nasdaq again looks narrow when Seagate sinks -9.2% and nuclear rips double digits on same day - classic late cycle rotation. But SpaceX $40B debt for Nvidia chips proves hyperscaler capex isn’t stopping. STI giving back Tuesday with OCBC -4% is overdone, SG banks oversold.

Don’t chase SpaceX +7.6% on Morgan Stanley note - cheapest space/AI call is peak hype with Musk back over $1T. Nvidia record and Nasdaq record on same day as 10Y and 30Y at 24-year highs is divergence that never ends well. DayOne filing Nasdaq IPO into this is bad timing.

SGX -7% on Friday with no announcement behind it is ugly price action even with new 10-share lot rule for DBS, OCBC, UOB, SGX from today. But US +29k jobs crushes October hike odds to 1-in-5 and Nasdaq record close means Fed pivot is real - global risk-on will lift SG after washout.

SGX -6% and 10Y hitting highest since 2002 is scary for beta, but Micron with 75%+ of 2027 already sold out proves this is a structural supercycle not cyclical. Memory names leading overnight while market bleeds is classic decoupling - long memory, short index.

Cooler PCE is dovish and helps October odds drop, but Micron’s record $54.2B and $61.5B guide wasn’t enough - stock closed flat. With 10Y still above 5.2% and targets from $1100 to $2000 already priced, this is classic sell-the-news after 224% YTD run.

Hot PCE re-accelerates and puts October hike back above 70% contradicting Williams dovish take ahead of Oct 28 meeting, yet Micron proves AI supercycle decoupling with 16 strategic agreements for $100B RPO and $22B customer deposits, sustaining pricing power through fiscal 2027 and justifying $1500 to $1625 consensus.

Anthropic revenue up 12-fold but still $42B net loss asking for $2T valuation target is pure bubble math. That’s WeWork-level delusion. Treasury yields at multi-decade highs will kill these insane multiples and gold cracked shows liquidity leaving. This IPO will top the AI bubble.

First up-week in three for Wall Street is just a bear market bounce. Microsoft led movers because everyone hid in mega-cap. Now Trump rejected Iran’s seven-day Hormuz offer, Brent higher again, US futures lower into Monday — classic risk-off setup. Oil shock will kill this fragile up-week fast.

30Y at 5.5% highest since 2004 and Williams calling another hike reasonable is peak fear. Meanwhile Anthropic just dropped $11.6B on Akamai compute and stock jumped 21% after hours. This tells you AI is still spending while money gets more expensive. When rates peak, this AI infra bid will explode higher.

10Y through 5% at a 19-year high on blowout PMI and October hike odds at 70% is the capitulation. This is peak rates. Nasdaq ending its record run is buyable fear. Meta pushing Muse to glasses and pendant at Connect gives market a new growth narrative to rotate into when yields top.

SanDisk getting its highest price target right before Meta Connect kicks off tonight is perfect timing. Meta needs massive storage for its Muse agent and wearables. Alibaba’s Apsara cooling off this morning doesn’t matter, memory is the bottleneck. Burry shorting chips here will get squeezed hard.

Meta’s Muse topping the App Store is bearish for Alphabet at 354.97, not bullish. If Intel, AMD and Arm rocketed Monday because agents need CPU, it means Google’s TPU advantage is irrelevant. OpenAI, Anthropic and Meta will flood the store and kill each other’s pricing power.