Goldman's Asian Currency Thesis Was Right—Except the Dollar Was Louder

Goldman Sachs went long on three Asian currencies for 2026: the Korean won, the Taiwanese dollar, and the Malaysian ringgit. The thesis was clean—AI exports, record current account surpluses, and foreign direct investment. The market delivered a different report card. All three closed the year lower against the U.S. dollar. The won lost 2.1 percent. The ringgit shed 1.7 percent. The Taiwanese dollar was the worst performer of the group, down 3.05 percent. This is not a story of a wrong call. It is a case study in how macro cycles override micro fundamentals—a lesson that applies directly to crypto markets where narrative-driven trading often ignores the liquidity tide. The Context: Goldman’s Narrative Framework Goldman’s call rested on a two-part framework: AI capital expenditure as a growth driver and energy supply shocks as a drag. The bank argued that economies deeply embedded in the AI chip supply chain—South Korea, Taiwan, and Malaysia—would see their currencies strengthen due to ballooning current account surpluses. South Korea’s surplus was projected to nearly double to $300 billion, or 13.9 percent of GDP. Taiwan’s surplus hit 25 percent of GDP. These are extreme numbers. In traditional forex models, such surpluses should push the currency higher. Meanwhile, energy-importing economies like Thailand, Indonesia, and the Philippines were expected to suffer. High oil prices would widen trade deficits, fuel inflation, and force central banks into a policy corner. The relative trade was obvious: long the AI exporters, short the energy importers. The Core: Where the Model Broke The flaw was not the AI thesis. It was the omission of a single, overriding variable: the U.S. dollar. In 2026, the dollar index rose nearly 3 percent, driven by a Federal Reserve that held rates higher for longer than markets expected. That single force suppressed every Asian currency, regardless of fundamentals. But the relative performance tells a deeper story. The weakest AI currency (Taiwanese dollar, -3.05%) still outperformed the strongest energy currency (Philippine peso, -4.48%). The spread is 1.43 percentage points. Goldman’s relative-value thesis worked. It just got buried by an absolute dollar rally. This mirrors a pattern I’ve seen repeatedly in crypto. In 2020, during the DeFi liquidity crunch, I watched Compound’s lending protocol nearly freeze as liquidity vanished. The fundamentals of the underlying protocols were sound—overcollateralized loans, transparent oracles—but the systemic factor (a sudden withdrawal cascade) overrode all micro-level signals. The market’s obsession with DeFi TVL missed the liquidity risk. Goldman’s AI export thesis missed dollar risk. Another twist: the Chinese yuan. It was the only Asian currency that rose against the dollar in 2026, gaining 3.32 percent. Goldman maintained its USD/CNY 6.50 forecast, implying further upside. But this is not a market-driven signal. China’s central bank deployed a multi-tool intervention package—reserve drawdowns, offshore bill issuance, and capital account controls. The yuan’s strength is a policy artifact, not a validation of the AI trade framework. The Contrarian Angle: The Relative Trade Paid, the Absolute Trade Lost Most retail traders chase absolute direction. They see a call on the won and buy the won. When it drops, they claim the analysis was wrong. But the professional play was always the relative pair: long KRW/THB, long TWD/PHP, long MYR/IDR. The data confirms these pairs moved in Goldman’s favor. The market noise from the dollar obscured the signal. Crypto traders make the same mistake. During the 2021 NFT mania, I ran a systematic floor-sweep strategy on CryptoPunks, acquiring 15 assets at 4.5 ETH average. The narrative was that Punks were a store of value. But when liquidity rotated to other chains, floor prices dropped 30 percent in a week. The absolute trade lost. However, the relative trade—long Punks versus short lesser NFT collections—still generated a 4x return over the cycle. The macro rotation (ETH dropping, Bitcoin dominance rising) crushed the absolute but left the relative intact. The takeaway is clear: in both forex and crypto, the macro anchor (dollar strength, Bitcoin dominance, stablecoin liquidity) is the first-order variable. The micro thesis (AI exports, NFT rarity) is second-order. You can be right on the second order and still lose money if you ignore the first. Goldman’s model also embedded a hidden cliff. The entire framework rests on AI capital expenditure staying intact. The bank explicitly noted that if AI spending slows, the differentiation between AI exporters and energy importers collapses. I’ve see this pattern before—in 2022, the Terra/Luna collapse was preceded by months of on-chain data showing the peg mechanism was unsustainable. I shorted LUNA through regulated futures with a 3x lever, strict stop-losses, and a $150,000 base. The profit was $450,000. The lesson: every narrative-driven trade must have a predefined exit if the macro premise shifts. The Takeaway: Chop Is for Positioning This is a sideways market. The dollar is not breaking down. AI capex is not collapsing. That means the relative-value trade remains viable. For crypto, the analog is clear: short energy-sensitive altcoins, long infrastructure projects with real fee revenue. But hedge the dollar risk—or in crypto terms, hedge the Bitcoin dominance risk. I bought the silence between the candlesticks. The market doesn't care about your thesis—it cares about liquidity. Ledger books don't lie. The Korean won’s current account surplus is real. But if the dollar keeps rising, those ledger books stay underwater. The only absolute hedge? Discipline. In 2017, I built an arbitrage script on Bancor that returned 22 percent in three weeks. The script worked because I defined the risk parameters first, not the profit targets. Goldman’s call was mathematically sound. But math without a dollar hedge is just a better-looking bet. Volatility is the tax on indecision. Pay the tax early, or pay it big later.