Market Overview — January 27, 2026

Summary

• This is a bull market year, and the logic is unchanged. Global fiscal irresponsibility, US financial deregulation and leverage, and the dollar term-premium carry trade are jointly creating “too much liquidity,” while an “insolvable” supply-demand gap in compute demand underpins corporate earnings. Political events—Greenland, the tariff ruling, the Minneapolis shooting—bring volatility, but “they do not necessarily bring direction”; the market ultimately trades on earnings and liquidity.
• A yen carry-trade unwind will not recreate the 2024 crash. The acute shock came from rate volatility, not rates themselves—“not rate itself.” Dollar-rate implied vol is falling, yen implied vol is not especially high, and there has been no 2024-style vol spike. Yen depreciation actually helps a trade funded in yen, and the carry trade is smaller than in 2024. There will be a moderate unwind, but no trend toward a shock to global equities. Michael Burry is short Nvidia and Oracle on a combination of compute-chip depreciation and carry-trade logic; Jin takes the opposite view.
• The 10-year Treasury is essentially at its peak. Breaking 4.25% will be difficult and 4.5% even harder: “If you touch the upside, I would absolutely see it as an opportunity to go long Treasuries.” The reasons are threefold: no inflation and very poor inflation expectations; the Fed will be dovish under any chair—if Trump does not choose Kevin Warsh, Jin expects one additional cut; and the labor market is weak, with big tech set to cut jobs over the next few months. Falling rate vol is allowing the short-funding/long-duration term-premium carry trade to gather momentum, capping the 10-year’s upside; at 4.0%, reverse the trade.
• Do not take the other side of a sovereign state. A true fiscal blowup requires monetary policy to be completely shackled by inflation, whereas today governments are “fiscally irresponsible while pretending to be monetarily responsible.” Forcing a sovereign into default “is like Chernobyl: it blows up at one point,” making timing nearly impossible while negative carry bleeds throughout the wait. The right trade is the opposite: fade rates whenever they reach the top of the range.
• The cloud and GPU earnings cycle is decisively bullish. The chip-depreciation short thesis was fully reversed in November and December: H100 rental rates surged because the new H200 and B100 cards were “completely unavailable,” and leasing agreements shifted entirely to long-term contracts. Higher memory prices have already left Xiaomi and Honor unable to ship phones. Microsoft reports earnings on the 29th; Azure growth is expected to accelerate to 39–40%, with capex roughly 100–125B, and intelligence points to a 1–2% beat. CPUs from Intel and AMD are also sold out across the board, but the earnings transmission will take several quarters—“it is a slow bull market, not a fast one.”
• The biggest near-term political risk is a US government shutdown. ICE has shot and killed a second person on the streets of Minneapolis, and Democrats are highly likely to use the January 31 expiration of the funding agreement to demand cuts to ICE funding and shut down the government; “the possibility is very high.” But a shutdown means debt issuance stops and supply freezes, sending 10-year yields rapidly down another 10–20 basis points. The hedge is bonds, not a flat book: “If you are flat, the compute revolution and demand for compute are still sitting there, alive or dead.”

Deep dive

1. The rates market has changed character: from risk hedge to a pricing tool for fiscal irresponsibility

• Jin’s opening framework is that stocks and bonds used to move inversely, with bonds serving as the risk hedge. But “over the past 18 months, the rates market has largely become a hedge against fiscal irresponsibility”—the trade is term-premium risk: investors see the US fiscal position as irresponsible and expect heavy issuance, so they sell bonds. This is the common key to understanding both the yen and dollar rates.
• Fiscal irresponsibility is a global, across-the-board phenomenon, and “fiscal irresponsibility may come with simultaneous monetary irresponsibility.” That is the pivot of the entire discussion: central banks will not truly stay hawkish, leaving liquidity excessively loose. Credit expansion plus deregulation forms the underlying fuel for the bull market Jin has been calling for this year.
• A quick mark-to-market: Jin recommended entering swaps when dollar rates were at last week’s highs. “If you went into swaps last week, you feel very good now”—rates subsequently crashed lower.

2. Japan: 高市 bets on an election, tax cuts stoke inflation

• 高市’s approval rating could rise above 70, but the LDP does not hold a majority in the lower house, so she may dissolve parliament and call an election while her support is high—possibly on February 9. The aim is a new fiscal package: cut the food consumption tax by ¥5T, taking the total fiscal deficit to ¥122T. The yen weakened and the JGB curve steepened that week; local banks cut duration while hedge funds unwound long-end positions, with both forces hitting the long end at once.
• Friday’s BOJ statement was expected to be hawkish, but it “deliberately” said yen depreciation could benefit inflation. The market read that as evidence the BOJ cannot truly stay hawkish. Jin’s signature view is that “the BOJ’s true allegiance is always on the side of preventing deflation（日本央行的屁股永远真正的是坐在防止通缩一边）.” The Fed, ECB, and PBOC are the same: they maintain the appearance of controlling inflation, but are actually more afraid of recession and deflation.
• Before the election, the BOJ conducted a bilateral inquiry through the New York Fed, which the market read as an intervention signal; dollar-yen crashed directly to 155 and 154. But Jin says the BOJ’s real intention is to let the exchange rate “fall steadily there.” Short-term intervention cannot change the simultaneous trend of a weaker yen and higher rates.
• A turn is already visible: large real-money accounts stepped back in to buy duration on Thursday and Friday, while 高市 softened her rhetoric. The government’s mention of bond buybacks—“not much different from directly printing money”—and extending the timeline for deficit implementation helped yen rates fall somewhat. A fiscal self-detonation is unlikely.

3. The carry-trade unwind thesis: vol, not rates, is the key; Burry is wrong to analogize to 2024

• The market is split. One camp sees little impact; the other, led by Michael Burry, is short Nvidia and Oracle on the basis of compute-center chip depreciation, with a carry-trade unwind as part of the thesis. Jin’s rebuttal has 3 layers: acute shocks come from rate volatility, not rates themselves; 2024 was a sudden vol spike, while dollar-rate implied vol has been falling and yen vol is not especially high; and when the yen weakens, “you borrowed yen,” so the move actually helps the carry trade. The trade is also smaller than Burry and others claim.
• The honest conclusion includes a hedge: a moderate unwind will occur, but “there is no trend toward a large-scale unwind causing a major shock to equities.” The carry trade is essentially a reservoir—Japan has printed money for years to supply liquidity to the world, especially US capital markets—but Jin does not expect it to create a major problem for global risk assets.
• The broader principle is that when taking the other side of a central bank or sovereign, “you have to keep pressing until it blows up, unless inflation runs out of control.” China’s M0 may have doubled, and its monetary base has doubled; much of the capital-market performance over the past 18 months was effectively printed. A true blowup is “like Chernobyl: it explodes at one point,” not over an identifiable cycle. That point is extremely difficult to predict, while negative carry bleeds throughout the wait—“not a particularly pleasant holding experience.”

4. The 10-year Treasury is nearing its peak: 4.25% is hard to break, 4.5% is a buy, and the term-premium carry trade is gathering momentum

• Three things are unlikely to change much this year. First, there is no inflation, and “inflation expectations are very poor.” Second, the Fed will be dovish regardless of which of the 4 candidates becomes chair—the Fed’s “train has already left the station.” It will become more market-oriented, cutting quickly when markets fall, while potentially hiking quickly when risk appetite overheats, as under Greenspan. Kevin Warsh is risk-off and has consistently advocated shrinking the balance sheet, but if Trump does not choose Warsh, Jin expects one additional rate cut. Third, the labor market is weak: banks cut 5% of their staff last year, tech companies will cut over the next few months, and “AI itself is a productivity technology replacing them.”
• Rates volatility has been trending lower for more than 6 months. When vol is low, the simplest trade is to borrow short and buy long-duration bonds—or borrow short-term money to buy mortgages. This term-premium carry trade in the US “has just begun; positions are not large, but it will continue” until a second inflation wave appears, which “may have to wait until next year.” Deregulated banks buying more Treasuries adds another major constraint on the 10-year’s upside.
• The instruction is blunt: at the upper end of 4.25–4.5%, “it is absolutely an opportunity to go long Treasuries.” Jin remains biased toward buying the long end. But the market is range-bound: “When it really gets to 4.0, you can reverse.” Fiscal irresponsibility does not mean shorting rates into a blowup; “the trade is the opposite.”

5. Compute shortages are insolvable: H100 rents reverse the short thesis, and cloud earnings are broadly bullish

• The core short thesis on cloud and Nvidia was that every new chip would make the old generation cheaper—“I buy a house and the rent keeps falling.” That was fully reversed in November and December: H100 rental rates surged because the new cards were impossible to rent. “If you want to rent a B100 now, you simply cannot get one.” H200 and B100 are completely unavailable, leasing agreements have shifted entirely to long-term lockups, and startups are scrambling back for H100s—“getting one at all is already good enough.” A recent JPMorgan report called the supply-demand imbalance insolvable.
• The shortage is spreading outward. Higher memory prices have left Xiaomi and Honor unable to ship phones. “Even cars cannot ship, because consumer electronics no longer count as customers” when capacity is being allocated first to compute centers.
• Microsoft reports earnings on January 29. Azure may accelerate to 39–40% growth, capex is roughly 100–125B, and “our intelligence is that it will beat, exceeding the market by 1–2%.” The logic is the same as a landlord’s: “rents are rising; I collect rent, and the rent is increasing.”
• Intel is a separate story. It was unwilling to invest in lagging-edge process capacity at the time, and demand was so strong that inventory was fully consumed while production could not keep up—“so the stock took a hit.” Jin does not think AMD taking share from Intel is a real factor: AMD is also supply-constrained, and CPUs are sold out across the board. The difference is timing: CPU shortages take several quarters to show up in earnings—“it is a slow bull market, not a fast one”—while the GPU shortage is “very, very certain.”

6. Politics creates volatility, not direction: a shutdown is highly likely and would favor long-duration bonds

• The big takeaway from Davos is Trump’s public drama over Greenland: the US is shifting “from a globalized America to a Monroe Doctrine America,” referring to NATO allies in the third person as “they.” The ideological fight between left and right has spread across the world. With the Supreme Court’s ruling on IEP A tariffs still pending, near-term uncertainty is only increasing.
• The domestic trigger is Minneapolis. ICE shot and killed someone in the street, and there is now a second fatality. Jin expects Democrats to go all in: when the latest stopgap funding agreement expires on January 31, they will likely use cuts to ICE funding as leverage to shut down the government—“the possibility is very high”—while pushing street protests and closing the government.
• The trading implication is counterintuitive. Once the government shuts down, debt issuance stops and supply stops, giving 10-year yields room to fall rapidly by another 10–20 basis points. The hedge is long bonds, not a flat book. Being flat is “very inappropriate,” because the compute revolution and demand for compute are still there. During the Black Lives Matter period, US equities rose because liquidity and corporate earnings were strong. “The situation is the same now.”
• The conclusion returns to the central thesis: this is “a trend that is moving upward; you may suffer a lot in the process.” The size of any drawdown and whether the market falls at all remain uncertain, but the market ultimately trades on earnings, economic growth, and liquidity. “The foundation is not bad, but the volatility along the way will be very large.”

Verification Notes

• The raw caption says Microsoft capex was “roughly 100–125B”; the currency and normalized amount cannot be confirmed.