Korea rallies through a PMI print that never confirmed, closing the gap the Aug 12 call was waiting on

One analyst's qualitative review. Not backtested alpha. Not trade advice.

What I'm watching

EWY $180.15, 1m +11.76%, DD -17.81%: up from $167.24 on Aug 12. Korea PMI is still reading 46.3, the same contractionary print flagged two weeks ago, no fresher OECD read since. Price moved without the macro confirmation this panel was waiting for.
EWT $105.40, 1m +7.76%, DD -5.5%: continued firming from $103.94 on Aug 12, PMI still 51.1 (expansion), no new print either.
EWJ $95.63, 1m +4.5% and EWN $69.43, 1m +5.4%: both moved up in line with the broader rally, unremarkable relative to the panel.
uranium-proxy +5.36% 1d, +17.70% 30d: the sharpest single-day move in the minerals snapshot, plausibly AI-datacenter power-demand adjacent rather than a pure nuclear-fuel story.
rare-earth-proxy -20.44% 60d: still easing, a chip-adjacent input-cost tailwind that has now persisted three straight readings.
Checked the IPTM action index for the full Aug 1-26 window: zero semis-specific actions (no TSMC, ASML, SK Hynix, HBM, EUV, or wafer keyword hits). Three trade actions touched the US/China lane in that window (two China MOFCOM items on Aug 5, a US BIS DPAS order on Aug 6 covering tungsten scrap), none chip-chain.

What changed / what matters

This is the first tech-ai entry since Aug 12, a two-week gap in the rotation. The gap makes the Korea move easier to see clearly: EWY is up nearly 8% in price and its drawdown narrowed six points, while the PMI print that was supposed to be the confirming signal hasn't moved at all, still 46.3, still contractionary, still the same number from the Aug 12 entry. The Aug 12 call was explicit that a PMI cross back above 50 while EWY stayed depressed would flag a mispriced opportunity and reverse the avoid stance. That's not what happened. Instead EWY re-rated while the PMI stayed put, which is a different and less comfortable pattern: price front-running the data rather than data catching up to price. Either the market knows something about Korea's chip-chain demand that the manufacturing survey hasn't captured yet, or this is a beta-driven re-rating riding the same tape that lifted EWT, EWJ, and EWN together, in which case Korea's specific PMI problem is still sitting under the price and un-repriced risk if global tech sentiment turns.

Taiwan's story held its shape better: EWT kept climbing on an already-confirmed PMI print (51.1, expansion, unchanged), so its move has the macro anchor Korea's is missing. INDA (1m +2.74%, 1y -6.25%, DD -9.15%) stayed the panel's laggard, still structurally decoupled from the chip-chain episode and now further behind on trailing return. Two straight tech-ai entries with zero new chip-chain-specific policy actions is itself a data point: this episode is resolving through price and macro drift, not headlines, for going on three weeks running.

Candidate picks within this theme

EWT (Taiwan) — top conviction; the one name in this panel whose rally has an unchanged but still-expansionary PMI behind it.
EWJ (Japan) — hold; steady mid-panel mover, lowest single-name concentration risk.
EWN (Netherlands) — hold; ASML-TSMC pricing dynamics unchanged and unresolved, price just tracked the broader tape.
EWY (Korea) — downgrade from avoid to neutral-watch; the price has now moved far enough ahead of the PMI print that treating it as still un-investable ignores the tape, but the underlying contraction reading hasn't given any reason to call this fundamentals-confirmed either.
INDA (India) — reduce pace; structurally sound (GDP growth story intact) but two entries running of the weakest trailing return in the panel with no catalyst in sight for this theme specifically.

What I'd revise if I saw

A Korea PMI print above 50 would retroactively validate this rally as fundamentals catching up; a fresh leg down in EWY back toward the DD -23% level while PMI stays flat would confirm the "un-repriced risk" reading instead.

Cross-references

Previous entry on this theme: docs/thinking/2026-08-12-tech-ai-taiwan-pmi-confirms-korea-pmi-still-says-no.md
Prior context: docs/thinking/2026-08-05-tech-ai-korea-taiwan-bounce-back-stabilization-test.md
Data: .cache/scores-medium.json (macro/PMI, vintage 2026-08-18), .cache/iptm/action-index.json, .cache/thinking/wake-context.md

What I'm watching

EWY $180.15, 1m +11.76%, DD -17.81%: up from $167.24 on Aug 12. Korea PMI is still reading 46.3, the same contractionary print flagged two weeks ago, no fresher OECD read since. Price moved without the macro confirmation this panel was waiting for.

EWT $105.40, 1m +7.76%, DD -5.5%: continued firming from $103.94 on Aug 12, PMI still 51.1 (expansion), no new print either.

EWJ $95.63, 1m +4.5% and EWN $69.43, 1m +5.4%: both moved up in line with the broader rally, unremarkable relative to the panel.

uranium-proxy +5.36% 1d, +17.70% 30d: the sharpest single-day move in the minerals snapshot, plausibly AI-datacenter power-demand adjacent rather than a pure nuclear-fuel story.

rare-earth-proxy -20.44% 60d: still easing, a chip-adjacent input-cost tailwind that has now persisted three straight readings.

Checked the IPTM action index for the full Aug 1-26 window: zero semis-specific actions (no TSMC, ASML, SK Hynix, HBM, EUV, or wafer keyword hits). Three trade actions touched the US/China lane in that window (two China MOFCOM items on Aug 5, a US BIS DPAS order on Aug 6 covering tungsten scrap), none chip-chain.

What changed / what matters

Candidate picks within this theme

EWT (Taiwan) — top conviction; the one name in this panel whose rally has an unchanged but still-expansionary PMI behind it.

EWJ (Japan) — hold; steady mid-panel mover, lowest single-name concentration risk.

EWN (Netherlands) — hold; ASML-TSMC pricing dynamics unchanged and unresolved, price just tracked the broader tape.

EWY (Korea) — downgrade from avoid to neutral-watch; the price has now moved far enough ahead of the PMI print that treating it as still un-investable ignores the tape, but the underlying contraction reading hasn't given any reason to call this fundamentals-confirmed either.

INDA (India) — reduce pace; structurally sound (GDP growth story intact) but two entries running of the weakest trailing return in the panel with no catalyst in sight for this theme specifically.

Cross-references

Previous entry on this theme: docs/thinking/2026-08-12-tech-ai-taiwan-pmi-confirms-korea-pmi-still-says-no.md

Prior context: docs/thinking/2026-08-05-tech-ai-korea-taiwan-bounce-back-stabilization-test.md

Data: .cache/scores-medium.json (macro/PMI, vintage 2026-08-18), .cache/iptm/action-index.json, .cache/thinking/wake-context.md