Goldman Sachs published a bruising pair of Asia notes over the weekend: strategists cut their 12-month KOSPI target to 9,000 from 12,000 after a 35% drawdown from June highs, even as they argued Korean equities remain deeply undervalued; separately, economists trimmed China's growth path and highlighted effective fiscal tightening, a 1.8-point property drag, and retail sales that barely moved despite stronger factory output.
Monday, October 26 opens with Goldman Sachs talking about Asia in two dialects. Equity strategists led by Timothy Moe cut their 12-month KOSPI target to 9,000 from 12,000, a 25% reduction published Sunday after Korean press picked up the note. China economists, in parallel weekend research, kept growth near Beijing’s floor but stressed fiscal drag, a property sector still subtracting roughly 1.8 percentage points from GDP, and consumption that refuses to catch up with factory lines. Both stories mention AI and semiconductors. Only one is mainly a global capital-markets repricing.
Seoul: memory beta at a discount
The KOSPI’s problem, in Goldman’s framing, is not a mystery recession in Korea. It is concentration and narrative. Samsung Electronics and SK hynix dominate index weight, and the benchmark has fallen more than 35% from its June peak as investors question whether hyperscaler AI infrastructure spending will earn its cost of capital. Financing rates, oil, and hawkish central banks amplified a chip-led repricing that already hit Taiwan and U.S. semis.
Goldman’s revision is conservative math on top of that shock: forward valuations on the index have collapsed toward four times earnings on some measures — among the lowest in Bloomberg’s 18-market set — with memory names allegedly pricing in a 65% earnings drop into 2027. Non-memory constituents, roughly half the index, trade near 7.6 times forward earnings. Foreign investors, especially hedge funds, kept selling; Goldman cited about $1 billion of outflows from Korean equities in recent flow data.
Yet the bank stayed constructive. The new 9,000 target still implies more than 60% upside from levels near 5,550 at the prior close. Strategists argued the market has baked in “excessive pessimism” and highlighted power, industrials, governance-reform beneficiaries, and K-culture domestic demand as places to hide from pure memory beta. That is a CAPITAL story: who is forced to sell liquid semi longs when prime-broker margin tightens, and who might buy when multiples imply cyclical extinction.
| Market | What hurts | What still works |
|---|---|---|
| South Korea | AI capex doubt × memory weight | Cheap non-memory, reform, domestic brands |
| China | Property × fiscal tightening | Exports, IP, selective stimulus |
| Link | Shared semi supply chain | Different household balance sheets |
Beijing: production without a shopper
China’s weekend packet points elsewhere. Q3 GDP printed 4.5% year on year, enough to keep the official 4.5–5.0% band within reach, and industrial production accelerated to 5.5% in September. Retail sales improved only to 1.1% — better than summer lows, not a consumption revival. Goldman economists described a widening production–consumption gap: factories hum on export and tech shipments while households face weak hiring, softer wage growth, and housing wealth still deflating.
Property research published the same weekend deepened that picture. Goldman estimates the sector’s direct drag on growth at 1.8 points in 2026, easing only gradually as the market’s share of GDP shrinks toward 8% from a peak near 30%. Land sales revenue fell 28% year on year through September; the shift from presales to completed-home sales tightens developer liquidity. Fiscal policy, meanwhile, effectively tightened in 2026: combined budget revenues grew slowly while Goldman lowered its augmented deficit forecast toward 10% of GDP from a higher prior — not austerity, but less impulse than bulls hoped.
The PBOC’s planned pivot toward an overnight policy rate next year would trim borrowing costs modestly. That is STATE architecture: incremental easing to defend the growth target, not a reopening boom. Culled’s recent China AI policy frame still applies — Beijing will export chips and models on its own timetable — but Sunday’s numbers show the domestic engine is property and local-government finance, not U.S. safety rhetoric.
Why the distinction matters for investors
Treating Korea and China as one “Asia risk-off” trade blurs the exit. Korea’s ask is whether AI memory earnings normalize before leveraged holders finish exiting — a question tied to U.S. hyperscaler capex, oil-driven rates, and duration fears in global bonds. China’s ask is whether fiscal and housing policy can stop bleeding household confidence without reigniting developer leverage — a slower, more political calendar.
Goldman did not contradict itself; it mapped two balance sheets. Seoul discounts a cyclical scare in the world’s most AI-exposed equity index. Beijing meets its growth target with exports and manufacturing while domestic demand underperforms. Same geography on the map, different fault lines underfoot.
What to watch
Three signals decide whether the weekend notes age well:
- KOSPI flows — Do foreign funds return if memory earnings revisions stabilize, or does governance and IPO supply (SK hynix’s Solidigm filing) keep pressure on multiples?
- China property prints — Land sales and completed-home policy are the lever behind Goldman’s 1.8-point drag; retail cannot recover while they slide.
- AI capex guidance — U.S. hyperscaler commentary remains the transmission belt into Korean memory; it does not fix Chinese apartments.
Until those diverge, “Asia” is not one trade — it is two stress tests sharing a wafer supply chain.
Continue reading
Sources
Goldman Sachs KOSPI strategy note via Seoul Economic Daily, Asia Business Daily, and CNBC (Oct. 25–26, 2026); Goldman Sachs China research summaries on Q3 GDP, fiscal deficit, and property drag (Oct. 25, 2026); Culled coverage of hedge-fund de-risking and China AI policy.
