Market Overview — February 3, 2026

Summary

• Trump has all but settled on Kevin Warsh to succeed Powell as Fed chair. But the host’s core judgment is that changing the chair will not reverse the broad liquidity-easing trend—reiterating December’s line that “the Fed’s train has already left the station.” The Fed operates under a 12-member voting system; post-2008 balance-sheet expansion was “everyone on board,” and “if you come in trying to flip the table, you still can’t get it done inside the Fed.” The rates market’s near-total lack of reaction is the proof: the retail-heavy markets—gold, silver, bitcoin and equities—were the ones “scared shitless,” with the impact increasing as retail participation rises.
• Trump has only one criterion for choosing a Fed chair: “Loyalty that isn’t absolute is absolute disloyalty.” Warsh’s father-in-law was Trump’s college classmate, a committed supporter and longtime donor—“he’s just picking one of his own kids.” Trump fundamentally wants a tidal flood of liquidity to support AI and the economic transition, so he cannot choose someone who will genuinely oppose him. Base case: Warsh talks QT but cannot deliver it, producing large-scale rate cuts or even QE, or uses bank deregulation to let private-sector leverage replace public-sector leverage—“that’s a tidal flood of liquidity.”
• The central bank is forced to hike only when inflation runs out of control, but the host believes inflation will not run out of control now. On services, “from tech companies down, everyone is laying off workers, and 700 million knowledge workers are going to be replaced by these models.” Shelter has an 18-month lease lag: “If you make a complete mess with a tidal flood of liquidity at year-end, it normally won’t show up until 2027.” On goods, China is “an enormous deflationary force”—its debt cycle “can only snap violently, like Chernobyl”—but the timing is impossible to predict and unlikely this year.
• Gold and silver’s January surge was a physical-market short squeeze. Trump threatened tariffs on precious metals, prompting silver to move from the LME to the US; physical shortages were compounded by ETFs locking up supply. Goldman’s regression shows the price impact of a 1,000-tonne purchase rising from 2% to 7%—“it used to be $30 moving the price by 2%; now it’s $120 moving it by 7%.” Last week’s crash came after the CME raised margin requirements and directly deleveraged the market, triggering a chain of long liquidations. He does not expect silver tariffs to be imposed in the end; going forward, “if it rises, it will be a volatile rise.”
• The dollar-yen carry trade will not bust. The key variable is volatility, not the level of interest rates; Japanese rates are stabilizing, and last week’s dollar weakness was caused by several Japanese threats to intervene. “I think it’s very difficult to make the carry trade unwind afterward.”
• Tech earnings are strong this reporting season, yet tech stocks are falling. SoFi’s earnings expectation, which sounded close to “sofa” in the remarks, was 7% but came in at 11%, while 2025 and 2026 earnings expectations are also strong. Hyperscaler capex consensus rose from $540B to $560B, implying nearly 38% growth over 2025; he expects the full-year figure to reach 40-50%, far above the 20-30% assumptions embedded in models. Hedge-fund positioning shows a sharp scissors divergence: persistent selling of software and buying of semiconductors.
• Oracle has not backed down despite being “shorted by the entire market.” It raised about $20B through stock issuance, taking total funding to roughly $50B, and continues to invest in capex. “These companies are now like the seven warring states: reform comes with growing pains, but if you don’t reform, you die.” Silicon Valley remains “very, very determined” to invest in AI while Wall Street is still questioning it; the Google vs. GPT and Grok-as-dark-horse model comparison will be pushed to next week.

Deep dive

1. Warsh Is Nearly a Done Deal, but “the Train Has Already Left the Station”

• The Fed did not cut rates at this meeting, but the host’s read of its message was that economic data are improving, the near-term case for a cut is insufficient, and committee members are aligned on a year-end downtrend in rates. The market interpreted “improving data” as hawkish, and the prospect of a more hawkish chair triggered the panic. His December view remains unchanged: “The Fed’s train has already left the station; whoever takes over will not make a particularly big difference.”
• The market’s cross-asset reaction is information in itself. Rates barely moved—if investors really believed Warsh would sell Treasuries and pursue quantitative tightening, “wouldn’t your rates have to shoot up?” Instead, gold, silver, bitcoin and equities were “scared shitless,” with the impact increasing as retail participation rises.
• Reiterating last week’s view, the dollar-yen carry trade will not bust. The relevant variable is volatility, not the size of the rate move; Japanese rates are stabilizing, and last week’s dollar selloff was driven by several Japanese threats to intervene. “I think it’s very difficult to make the carry trade unwind afterward.”

2. Trump Wants a Tidal Flood of Liquidity—Base Case for a Hawkish Chair Is QT Turning into QE

• Warsh made his name after 2008 by sharply criticizing Bernanke’s balance-sheet expansion: “The Fed shouldn’t bail out the market; all inflation is caused by balance-sheet expansion.” The host’s rebuttal is equally blunt: “He’s not wrong, but without balance-sheet expansion and monetary-policy support, we would have been in deflation—that’s the part he doesn’t discuss.” Giving policy advice from outside government and actually governing are different things; once in office, he would still have to respond to employment and economic growth.
• The institutional constraint is hard. The Fed is governed by a voting system—it is “not a one-leader system like China, where the central-bank governor can double the monetary base simply by deciding to.” Powell had to use his credentials to lobby members one by one, and even when he pushed through a December cut, there were already dissents on the committee. “It is actually difficult for Warsh to operate outside the Fed’s institutional framework and the fundamentals of the US economy.”
• Trump’s selection logic has only one rule: “Loyalty that isn’t absolute is absolute disloyalty”—policy “must ultimately follow the leader’s will.” The host pointed to the pre- and post-election reversal on China policy, with the name rendered as “Lieutenant” in the remarks, apparently referring to Lighthizer, as well as Secretary of State Rubio. Warsh’s father-in-law was Trump’s college classmate, a committed supporter and longtime donor: “He’s just picking one of his own kids.” Warsh can also clear the Senate—“it’s easy for a very dovish candidate to get blocked in the chamber, so choosing Warsh is a fairly smart move.” Powell, meanwhile, faced pressure from Trump through a criminal investigation tied to renovations of the Fed building after refusing to cooperate on rate cuts.
• The real difficulty with QT is that it would reverse decades of work after 2008 replacing private-sector leverage with public-sector leverage. He recalled the high-leverage banking era, when “my boss’s bonus was $50 million.” To sustain US growth above 2.5%, neither population nor productivity is sufficient: “You have to print money to get there.” If Warsh genuinely shrinks the public sector, he would have to unshackle banks and remove too-big-to-fail-style constraints so private leverage can take its place. “You could end up with public-sector leverage not declining, private-sector leverage rising, and rates falling again—that’s a tidal flood of liquidity.”

3. The Fed Has Never Been Independent, and Inflation Will Not Run Away

• “The Fed is never independent—people say it’s independent, but it never has been.” No fiscal policy anywhere can pursue irresponsible behavior independently of monetary policy. When Chinese local governments spend aggressively, if the central bank withholds funding, “every local government going bankrupt becomes inevitable.” The only historical attempt not to coordinate was the European debt crisis; the cost was Europe’s decline, a disastrous economic record over the past decade and a clear widening of the gap with the US—“the next 10 years will be even worse.”
• Central banks are forced to hike only when inflation runs out of control, as in 2021. Inflation is “very lagging,” and none of its 3 major components is currently capable of pushing it higher. On services, there is no Biden-era labor shortage; “from tech companies down, everyone is laying off workers, and 700 million knowledge workers are going to be replaced by these models.” Shelter has an 18-month lease lag: “Even if you create a complete mess with a tidal flood of liquidity now, it normally won’t show up until 2027.”
• China is “an enormous deflationary force” on the goods side: products are cheap, labor is intensely competitive, and the unemployed are “all laborers helping bring down your inflation.” The hedge remains unchanged: China’s debt cycle “can only snap violently—prices suddenly surge while wages do not, like Chernobyl—but the timing is difficult to predict, and you are unlikely to produce that outcome this year.” Politically, the Treasury secretary is probably Bessent—the name sounded like “Bilson/Bellson” in the remarks—and apparently handles China negotiations as well as a range of other matters; he “has Trump’s ear.” “I find it difficult to believe he could support someone completely unlike him as Fed chair.” Ultimately, “I don’t think Kevin Warsh has much to worry about.”

4. Gold and Silver: A Physical Squeeze Drove the Surge; Margin Deleveraging Drove the Crash

• The surge was driven by a physical short squeeze. Trump threatened tariffs on precious metals in January, prompting the market to worry that silver moved into the US would be taxed. Physical silver was pulled from the LME into the US, leaving insufficient LME spot supply while ETFs kept holdings locked up. Goldman’s linear regression quantified the fragility: the price impact of a 1,000-tonne purchase rose from roughly 2% to 7%—“it used to be $30 moving the price by 2%; now it’s $120 moving the price by 7%.” ETFs and leverage compounded the move, making silver extremely expensive. Gold rallied as well, with the PBOC buying consistently.
• The direct mechanism behind the crash was deleveraging. The CME raised margin requirements and “directly deleveraged” the market; combined with rumors of the extremely hawkish Warsh appointment, that “blew out the longs all the way down.” There was already some rebound last night, but the market needs to wait for the clearance process and liquidation of the remaining positions to finish.
• The forward view is conditional. “It is difficult for Trump to actually put silver under a tariff,” so he does not expect the market to continue along the same path. Silver has genuine demand from uses including solar power, while gold requires asking what the PBOC plans to do. Using storage as an analogy, there is a real shortage—so severe that shipments of second- and third-tier phones might have to be cut by 40-50%, and even cars cannot ship—but capacity is also expanding. “If it rises, it will be a volatile rise,” not a straight line with no volatility.

5. Earnings Season: Capex Only Goes Higher; Oracle Says Reform or Die

• Tech earnings are broadly good, yet the stocks are falling. SoFi—the name sounded close to “sofa” in the remarks—was expected to post 7% earnings growth but delivered 11%, and earnings expectations for 2025 and 2026 are also strong. “Earnings growth in tech and semiconductors is still very certain this year.” But hedge-fund positioning in software versus semiconductors shows a huge scissors divergence: software keeps being sold while hardware keeps being bought. What happens to the software ecosystem is the question he is reserving for next week’s model-focused discussion.
• Hyperscaler capex consensus rose after earnings from $540B to $560B, implying roughly 38% growth over 2025. He believes the full-year figure will reach 40-50%, while every financial model is still built around 20-30% growth assumptions.
• Oracle is the extreme case: “the entire market is short it,” betting that capex is excessive, financing is too aggressive, cash flow will not return until 2028, and its eventual customer base is just OpenAI. It nevertheless raised about $20B through stock issuance, bringing total funding to roughly $50B. “Even if it sells what it has, it is going to invest in capex and AI.” His conclusion captures the whole debate: “These companies are now like the seven warring states. Reform comes with growing pains, but if you don’t reform, you die.” Silicon Valley is “very, very determined,” while Wall Street still does not really believe in AI in the short term. “But there’s no way around it. As a company, you have to overcome that.”