Market Overview June 3, 2025
Summary
The May rally was built on the “TACO” consensus: markets decided that “Trump Always Chickens Out,” and stopped seriously pricing every tariff threat. Markets repriced for easing U.S.-China tensions, improving U.S. growth expectations, and future fiscal spending, pushing stocks higher and steepening the curve, with the 30-year yield near 5% and the 10-year near 4.5%; policy reversals remain the tail risk markets must price this year.
Two weeks ago, light positioning could still support the rally; that logic has now largely disappeared. Global CTA, risk parity, and vol-control strategies bought roughly $100B of stocks over the past month; over the next week and even the next month, a rising or range-bound market would generate almost no incremental buying, while a roughly 3-standard-deviation decline would turn systematic strategies into a selling “accelerator.”
Corporate buybacks will also enter their blackout window around June 16 to July 25, weakening another source of near-term liquidity support. Buyback volumes are already declining, with financials likely to stop buying first and tech next; positioning is not yet as dangerous as it was at the end of 2024 or the start of 2025, but it is “less supportive,” so marginal position-building can no longer explain further index upside.
The S&P 500 trades at roughly 25-26x current-year earnings, or more than 22x even assuming 10% growth this year, while long-end yields make that valuation harder to justify. During the pre-pandemic bull market, the S&P traded at about 23x earnings with the 30-year Treasury yield at only around 2.2%; with that yield now near 5%, the speaker sees broad-market valuations as elevated and recommends reducing overall exposure in an environment that could deliver either a soft landing or a one-quarter recession.
An expensive broad market does not mean the AI complex is expensive too; cloud and semiconductor names are instead areas where the speaker is willing to retain exposure. Meta, Google, Amazon, and Microsoft are reporting solid token utilization, and AI capex ROI remains healthy; Microsoft is first cited at roughly 30x earnings and later at perhaps 20x, versus around 22x for NVIDIA, 20x for Alibaba, and 50x for Costco—a disconnect in which earnings visibility is higher but P/E multiples are lower.
Tariffs have yet to hit the hard data: before fiscal support reaches households, the U.S. economy must face an effective tariff increase of more than 10% with no offset. The speaker estimates the economic impact of tariffs at 1%-1.5%, implying 2025 GDP could come in below 1%; a one-quarter recession could arrive in the fourth quarter, but is more likely to be delayed until the first quarter of next year, while even if policy is enacted in November, refunds and other cash transfers would not reach household wallets until March.
A court ruling will not eliminate tariff risk; the decisive variable remains Trump’s political will. The speaker believes the president can route around the courts through Section 301, Section 122, Section 232, or Section 102; Trump needs to satisfy MAGA, while China’s bureaucratic system rewards hawkishness and punishes dovish mistakes, so agreements may be reached but will be “fought over back and forth,” making sustained improvement in U.S.-China relations unlikely.
Tesla remains a company the speaker likes but does not want to own heavily, because the Robotaxi story has yet to fill the hole in automotive earnings. Around $450 once implied 20% growth this year and a June Robotaxi launch, but he believes the former is now unlikely and the market still needs to absorb the shift from positive 20% growth to negative 10%; roughly 50% excess capacity across China’s auto industry, Musk’s political positioning hurting core buyers in Europe and the U.S., and high long-term rates all suggest the selloff is not yet over.
Deep dive
1. Support for the “TACO” Rally Is Weakening
After the Switzerland meeting, the market’s most popular trade was summarized as “TACO”—“Trump Always Chickens Out.” The courts first struck down the tariffs, then an appeals court reinstated them, while Trump accused China on Truth Social of failing to honor its commitments. Investors nevertheless continue to treat these statements as reversible policy threats and are not pricing them seriously.
The May rally spanned both equities and rates: stocks rose, the yield curve steepened, and the 30-year yield reached roughly 5%, with the 10-year around 4.5%. Markets repriced around two assumptions—that Trump’s tariffs would not remain in place for long and that the U.S. government would support growth through fiscal spending.
Two weeks ago, the speaker saw “high valuations, unclear fundamentals, but very light positioning”—a different fragility profile from the end of 2024, when fundamentals and sentiment were strong but positioning was extremely one-sided. Global CTA, risk parity, and vol-control strategies have now bought roughly $100B of stocks in a month. Positioning is not yet as dangerous as it was at the end of 2024 or the start of 2025, but it can no longer provide a sustained marginal bid.
The flow profile ahead is asymmetric: in a rising or range-bound market, CTAs would largely sit still and generate almost no new systematic buying; a roughly 3-standard-deviation drawdown, however, would trigger substantial selling. Buybacks have long been an important source of net demand for U.S. equities, but volumes are also declining and will enter a blackout window around June 16 to July 25; financials may stop first, followed by tech. The logic of relying on positioning and buybacks to keep the market rising has therefore largely disappeared.
2. The Index Is Expensive; the AI Complex Is Relatively Cheap
The S&P 500 trades at roughly 25-26x this year’s earnings, or more than 22x even assuming 10% growth this year. Before the pandemic, when the market was in a bull market with no recession expected, the multiple was around 23x—but the 30-year Treasury yield was then about 2.2%, putting the corresponding risk premium at a “completely different level” from today.
The speaker’s conclusion is not to liquidate everything, but to reduce broad-market exposure and reallocate: financials and retail are particularly expensive, while semiconductors, cloud companies, and AI-related names are relatively cheap. If he were to retain exposure, he would favor semiconductors and cloud, even though NVIDIA still faces slowing growth and declining margins.
The case for AI is not conceptual; it is usage. Meta and Google use more of their capacity internally, while token utilization at the four major cloud providers is solid. The key variable in the AI trade over the past two years has been cloud capex, and those investments are in fact still being used, with capex ROI remaining healthy. Because the visibility on investment, revenue, and returns is relatively high, these companies may trade at lower P/E multiples than the broader market.
Valuations look somewhat “funky”: Microsoft has been among the stronger recent recoveries, first cited on the program at roughly 30x earnings and later, on another basis, at perhaps 20x; NVIDIA is around 22x, Alibaba around 20x, while Costco is near 50x. AVGO, Marvell, and the four major cloud providers were also cited as relatively attractive names.
3. Soft-Data Optimism Is Arriving Before the Hard Tariff Shock
Consumer confidence was strong last week, reinforcing the trade that “tariffs have been introduced and the economy is fine.” The speaker explicitly rejects extrapolating that directly into growth: soft data has had a poor forecasting record over the past two years, while the post-pandemic economic structure and government spending may further weaken its explanatory power.
The more important issue is the transmission lag. Tariff effects need at least 3 months, and possibly 6 months, to enter the hard data; current releases may still reflect an economy from before Liberation Day. The indicators that matter are nonfarm payrolls, unemployment, and similar measures—not immediate sentiment surveys.
Restocking and front-loading can also create a short-lived false boom. China’s external trade may be strong simply because companies shipped goods early in anticipation of higher tariffs. Tariffs can distort the quarterly path, but they do not eliminate the ultimate shock: “By year-end, whatever the impact should be will ultimately show up.”
The speaker estimates that an effective tariff increase of more than 10% would have an economic impact of roughly 1%-1.5%, implying 2025 GDP could come in below 1%. That shock could produce a recession within a quarter, or might not; his base case is that a recession is likely in some quarter, with the fourth quarter possible but the first quarter of next year more likely.
4. Global Fiscal Expansion Is Real, but the Market Is Trading the Timeline Too Early
Markets believe the U.S. will avoid austerity and stimulate demand through measures including tax-free tips, individual tax cuts, and refunds. The problem is timing: even if the refund policy is enacted around November, the cash would not reach household wallets until March of next year. Until then, the U.S. economy is effectively absorbing the full force of the higher effective tariff.
The larger bullish narrative is a global fiscal-spending cycle: the U.S. is preparing to expand, Germany and Europe are increasing military spending and investment, and Japan and China are also spending. The speaker acknowledges that the logic “makes some sense,” citing Chinese asset prices after 2008 as an analogy; his objection is timing—the optimism has been priced in too early, and a correction may still occur in the interim.
5. Courts Cannot Eliminate Tariff Risk; Political Incentives Will Keep Reviving the Conflict
The speaker viewed a Supreme Court ruling as positive for markets: without just cause, Trump cannot arbitrarily replace Fed officials. Separately, the Court of International Trade temporarily blocked his tariffs before an appeals court reinstated them, making the legal process itself a source of policy volatility.
The speaker found a Goldman Sachs report persuasive: even if the appeals court had not restored the original tariffs, the president could still attempt to impose new ones under Section 301, Section 122, Section 232, or Section 102. His summary was simple: “If the president wants tariffs, he can always impose tariffs.” In his view, neither the legal process nor Congress can currently stop the president from imposing tariffs in any meaningful way.
China’s mechanism is not simply a matter of negotiating tactics; it is bureaucratic incentive design. The speaker’s blunt formulation was: “When doves make a mistake, they get purged; when hawks make a mistake, they don’t.” Every compromise therefore invites attacks from the hardliners, pushing policy toward greater hawkishness and potentially weakening implementation of any agreement.
The U.S. faces political constraints as well. Trump may not put economic optimization first; he needs MAGA votes and standing within the GOP. The two sides may “somehow reach a few agreements,” only to see tariff threats against China or Europe resume. That is why the speaker has consistently defined 2025 as a “year of tail risk.”
6. Tesla Needs Earnings to Take Over; the Robotaxi Story Is Not Enough
The speaker emphasized that he has long been a loyal Tesla fan and that several profitable trades last year came from Tesla. He repeatedly said he would buy at $220, but did not act when the stock actually reached that level. Elon Musk’s return to Tesla satisfies one prerequisite, but elevated broad-market valuations make a “story-stock” particularly vulnerable when multiples are compressed.
A price around $450 once implied two expectations: 20% growth in 2025 and a Robotaxi launch in June. The first is now unlikely; the Robotaxi may appear in June, but it is unlikely to close the automotive earnings gap. The market can temporarily support the stock with a new story, but once that story is exhausted, it needs the next one to sustain the price.
Automotive fundamentals face three pressures. China’s automakers collectively have roughly 50% excess capacity, forcing Tesla—with around 20% market share in China—into an extreme competitive shakeout; excess capacity and parts competition could also spill into overseas markets. Musk’s political alignment will hurt left-leaning electric-vehicle buyers in Europe and the U.S., while higher long-term rates will suppress demand through auto loans.
The speaker believes the market still has to absorb the shift from positive 20% growth to negative 10%. “Once the story is over, it can’t support the stock anymore; you have to tell the next story.” Hardware 5 is not coming quickly, and takeoff may not arrive until the end of this year or next year. He still likes and follows the company and may eventually build a large position, but his current view is that the selloff is not yet over.