Summary

The speaker is sticking with his early-year framework: the U.S. is the main downside tail risk in 2025, while China and Europe are mainly upside “spike” tail risks that could also turn into risks during a U.S. crisis. A U.S. recession is still not the base case, but tariffs, reduced immigration, DOGE and potential problems in Europe will surface one by one; China and Europe’s upside “is not a trend”—if you missed it, don’t chase it.
The fuel for the short-term rebound is passive buying, not the disappearance of macro risk. CTA is short about $20B, vol control has largely liquidated, corporate buybacks are up roughly 10% year on year, and another $40B of quarter-end rebalancing demand is due over the next two weeks; “the main question is who is still going to buy blindly,” and these flows make another sharp near-term selloff harder.
The real tactical risk window is April, when ugly March data will be released in concentrated fashion. Reduced government software-seat usage and overbooking, companies halting expansion, and a falling stock market suppressing hiring could jointly weigh on government and private-sector spending, payrolls and consumption; but the speaker can still “see poor economic data in March, but not a U.S. recession,” and genuine recession pricing may require the index to fall to 4800 or below.
The April 2 tariff package could push the overall effective rate from roughly 4% today toward 10%, while squeezing both growth and the Fed’s room to support markets. The speaker cites estimates that GDP growth would fall from 2.2% to 1.7% and inflation would rise by about 0.8 percentage points; Powell is more likely to “stay out of this shit,” and while there may be two rate cuts this year, large-scale easing will be difficult.
The market’s biggest institutional concern is that the Trump Put is absent this year, while the Powell Put may not exist either. The speaker believes Trump may even want the stock market to fall so he can blame Biden; reciprocal tariffs could include subsidies and other “unfair” provisions, while imported cars worldwide could face a uniform 25% tariff, with retaliation turning negotiations into “a back-and-forth game of ping-pong”; a continued rally after the bad news is priced in remains possible, but investors without a long-term view should avoid the April risk.
Germany and China’s policy impulses can support temporary decoupling, but cannot withstand a genuine U.S. deleveraging. Germany plans to add roughly 1% of GDP in defense spending, launch 500 billion in infrastructure spending and push its deficit to about 3.5%, while China is also signaling fiscal and monetary expansion; but the speaker’s core judgment is that both are still “using fiscal policy to chase deflation,” and a U.S. crisis would pull dollar liquidity back and sweep global assets lower together.
Intel is one of the speaker’s few single-stock ideas with a long-term re-rating case, while whether the S&P 500 can return to 5800–6000 still depends on EPS. He believes 18A was already relatively mature last year, with decent yields; valuing the product business at 15x implies roughly $200B, or $40–50 per share. At the index level, if fiscal spending, AI investment and economic growth—the three pillars—hold, roughly 8% growth this year remains achievable; in a recession, EPS could flatten or turn negative, producing “a hit to earnings on one side and valuation on the other.”

Deep dive

1. 2025 is a year of policy tail risks, with U.S. downside and China/Europe upside

The speaker reiterated the framework he laid out on January 7: “It may happen, it may not,” but the direction of the tail risks is clear—the U.S. skews downside, while China and Europe skew upside. A U.S. recession remains the base case against, but the market reaction could be far larger than the change in fundamentals.
The four factors are tariffs, reduced immigration, DOGE—budget cuts and deficit reduction—and potential problems in Europe. DOGE has delivered limited actual deficit reduction so far; its political and ideological impact is much larger. But the government’s willingness to spend and overbook has declined, already reducing software seats and making the companies involved reluctant to keep expanding and hiring.
2023 traded AI, 2024 traded the U.S. avoiding a hard landing, and 2025 is a “policy year.” Policy uncertainty and liquidity are pushing capital toward Europe and China, but that is “an upside spike, not a trend”; the speaker’s rule is simple: “If you didn’t catch it, you didn’t catch it.”

2. Positioning has pushed short-term selling pressure to the limit, extending the rebound window through the end of March

This selloff has lasted nearly 20 days, with sentiment flipping rapidly from the post-election peak to negative. It has not reached the extremes of 2008–09 pessimism, but the speed of the decline has built considerable rebound momentum. The rebound only truly began on Friday, so it may last longer than initially expected.
CTA is short about $20B, making further shorting unlikely; vol control has largely completed its liquidation. The speaker is not focused on the 20-day or 50-day moving averages, but on “who is still going to be forced to sell.” The mechanical selling pressure from both strategies is now close to exhausted.
Corporate buyback flows are roughly 10% higher than a year ago. Though they will be constrained by the blackout period, Asset Managers have still been net sellers over the past 15 days, while Hedge Funds have clearly started covering. More importantly, quarter-end pension rebalancing is expected to force roughly $40B of stock purchases over the next two weeks, “regardless of price.”
Dealer gamma is currently near zero, meaning market and bank positioning is broadly Neutral and does not require corresponding hedging as prices rise or fall. On a static basis, another roughly 4% decline in the index could instead move the market into positive gamma. Combined with VIX falling on down days and some investors returning to short vol, the speaker believes “a continued plunge is very difficult” in the near term. That does not confirm a long-term bottom.

3. March data will look ugly, but recession remains unconfirmed

CPI and PPI were weak last week, followed by retail sales below expectations, pointing to further deterioration in private consumption and economic data. The speaker’s high-probability call is that “March economic data will look very ugly,” reviving recession trading when the figures are released in April.
The transmission mechanism for DOGE can be seen in ServiceNow and SNOW. A government department might once have had 10 people ordering 20 software seats; now, with employees idle and disengaged, overbooking is also contracting, weakening the chain through which government spending becomes private-sector earnings. Government spending may not fall sharply immediately, but the willingness to spend and overbook has already declined. Even without immediate layoffs, the companies involved are unlikely to keep expanding and hiring.
A falling stock market will further suppress hiring and payrolls. But the speaker distinguishes bad data from recession: “I’m not fully convinced of that claim yet.” Rates markets are currently pricing roughly 65bp of cuts, fewer than 3 reductions. To say the market is truly pricing a recession, he believes the index may need to reach 4800 or below.
He also believes DOGE, the Taylor Rule and the other factors above would need to align before there is greater confidence in calling a U.S. recession. It is still too early.

4. The April 2 tariff package will squeeze both growth and the Fed’s room to respond

The key import tariffs already imposed amount to an effective rate increase of roughly 4%. The April 2 package could exceed the original 4%–5% expectation and approach 10%. The risk is that reciprocal tariffs may match not only nominal tariff rates but also count subsidies and other “unfair” provisions; imported cars worldwide could also face a uniform 25% tariff.
The second-layer risk is retaliation. The speaker believes the U.S. is relatively insensitive to Chinese retaliation, but its industrial links with Europe are much deeper. If both sides raise tariffs, retaliate, negotiate and adjust rates, markets could endure a painful process in which the measures are “batted back and forth like a game of ping-pong.”
Under the scenario he cites, a 10% effective tariff could cut roughly 0.5 percentage points from GDP growth, taking it from 2.2% to 1.7%, while adding about 0.8 percentage points to inflation. Without the measures, year-end inflation might have been around 2.1%; now it could approach 3%, before accounting for immigration.
This makes it difficult for the Fed to rescue both growth and asset prices. The speaker expects two rate cuts may indeed occur this year, but large-scale easing is not on the table. Powell is more likely to “stay out of this shit.” The essence of the market decline is that “the Trump Put is not there, and the Powell Put may not be there either.”
The April 2 tariff package, combined with March economic data, creates a risk window to avoid. A short-term rebound could still continue after the bad news is priced in, but investors without a long-term view—and who do not regard this as a bottom—should prioritize controlling April risk.

5. Europe and China’s fiscal impulses can create spikes, but cannot promise a trend

Germany’s plan includes additional defense spending of roughly 1% of GDP and 500 billion in infrastructure investment, including 100 billion for climate initiatives, while pushing the deficit ratio from near zero to about 3.5%. European defense stocks and assets rallied first, but the speaker warned that “the market’s first reaction is often wrong.”
His view of the mechanism is that Europe is still “using fiscal policy to chase deflation.” The Fed can expand directly against Treasuries and, in theory, continue buying them; the ECB faces the bonds of individual member states and fiscal conditions—the speaker cites a requirement that deficits remain below 5%. German fiscal expansion will lift Bund yields and squeeze the financing capacity of high-deficit peripheral countries, creating a fuse through which U.S. risks could trigger European debt problems.
China is likewise signaling consumption stimulus and fiscal and monetary expansion, but CPI remains negative, while property, balance-sheet, demographic and private-income problems remain unresolved. The speaker emphasizes that no policy has genuinely raised private-sector income or repaired private-sector balance sheets, leaving consumption without meaningful expansion.
China and Europe can decouple for a time if the U.S. declines mildly or achieves a soft landing; the more severe the U.S. problem, the sooner that decoupling ends. The speaker’s conclusion is that “once deleveraging starts, liquidity inevitably flows back to the U.S.,” sweeping Europe, China and even Japan into the decline.

6. Intel’s re-rating bet rests on 18A, 陈立武 and U.S. manufacturing

On Intel, the speaker believes 18A was already relatively mature last year, with decent yields, and that the market has not fully reflected this. His private estimate is that if the fabs were spun off and only the product business were valued at 15x earnings, it would be worth roughly $200B, or $40–50 per share.
After 陈立武 (Lip-Bu Tan) became CEO, the speaker expects him to cut excessive and wasteful investment in AI while continuing to prioritize both the wafer and product businesses. Intel’s wafer business has already absorbed more than $100B of cumulative investment.
The longer-term policy logic is that the Trump administration’s key import tariffs will inevitably cover chips and require TSMC to move large-scale manufacturing to the U.S. The speaker compares Intel’s current position with AMD’s when it began using the x86 architecture again, arguing that TSMC must also help lift Intel back up. This “long slope, deep snow” story is about ready to get underway.

7. The S&P 500’s outcome depends on EPS; full reallocation awaits recession or mispricing

The speaker believes this U.S. equity cycle differs from 2000: the rally has been supported by both multiple expansion and earnings, with EPS growth potentially around 13%–14% in 2025 and 2026. What truly determines the full-year outcome is whether a recession causes earnings growth to flatten or turn negative.
Earnings rest on three pillars: government fiscal spending, AI-led investment and the underlying growth of the U.S. economy. The four major cloud providers have not cut AI investment, and DOGE has not yet genuinely shaken the fiscal cycle. If all three pillars hold, roughly 8% growth this year remains possible, making a return to 5800–6000 “very reasonable.”
The downside scenario is simultaneous contraction in EPS and valuation multiples—“a hit to earnings on one side and valuation on the other.” April is therefore not a preordained selloff, but a key point for reassessing recession odds based on March data, tariff details and market sentiment.
Fully rebuilding the position requires one of two conditions: a genuine U.S. recession that triggers global deleveraging and creates a “pretty frightening” buying opportunity; or a market that has begun pricing in recession even though the speaker believes the U.S. will not enter one. The latter window may emerge in April or May. Until then, short-term rebounds and long-term allocation should be treated separately.

Alex Wiltschko - Giving Computers A Sense Of Smell - [Invest Like the Best, EP.415]A Conversation with 马毅: Intelligence, DNA, and Entropy Reduction