US equities surged during the second quarter, with the Nasdaq climbing 32% and the Philadelphia Semiconductor Index skyrocketing 87%. But the market has since stalled, with that powerful rally grinding to a halt as stocks enter a sideways consolidation phase.
Goldman Sachs partner Mark Wilson notes in his latest Weekly Mash that American stocks have been rangebound since the Q2 surge, with three key headwinds suppressing the tape: the historical pattern of midterm elections, quant machines still running nearly full long, and a tense energy-bond-equity triangle. Wilson is simultaneously bullish on four specific themes: AI infrastructure leaders, commodities and related equities (especially copper), the German stock market, and large-cap banks.
Sizing Up the Three Key Obstacles
Obstacle One: The Haunting Specter of Midterm Elections. Wilson points to data going back to 1974, which shows the median return for US equities from early August through the midterms is exactly 0%. He admits no one truly knows how strong this pattern really is, but his read is straightforward: "The number is quoted so often that the market is now trading on it."
Obstacle Two: Smart Money is Out, But the Machines Are Still In. Hedge funds spent the summer de-risking, with gross exposure at mid-levels of the past year while stock-picking funds are even lighter. Wilson argues, however, this doesn't mean much for the index itself, since hedge funds typically dump both winners and losers when they pull back. The real threat lies elsewhere, with trend-following funds and volatility-targeting strategies still close to max long. "These flows don't hedge themselves. They simply sell when prices fall," he writes.
Obstacle Three: The Energy-Bond-Stock Stranglehold. Wilson calls this "the most tightly connected core contradiction in the market right now," working through a clear chain: geopolitics push oil and gas prices up, which lifts bond yields, which in turn pressures equity multiples. This is "the toughest puzzle in the book," he says, and yields have already traveled a long way. The latest development this week: Trump indicated the Iran issue will persist beyond the midterms, which Wilson interprets as "a precondition for the parties to start negotiating like adults." Oman and Gulf states are now set to discuss Strait of Hormuz access with Tehran. Goldman Sachs has also raised its oil and gas price forecasts again. On natural gas, Wilson notes "the price level that forces Asia to cut consumption and frees up supply for Europe has been basically reached," with prices breaking above the March high. Still, Europe will head into winter with historically low storage levels. Wilson stresses the first two hurdles will either erode over time or clear with a price pullback, but the energy-yield knot "won't expire in November."
Four Directions for Positioning
Direction One: AI Infrastructure - The Story Isn't Over, But Valuation Has Gotten Ahead of Itself. Wilson feels that looking back this late summer could reveal another inflection point: more models are emerging, GrokBot went viral, and the "deliverable AGI" timeline has sped up again. But he's less interested in the AGI narrative and more focused on capital expenditure. "Fundamentals continue to strengthen," he says, pointing to a slew of data: Oracle's latest earnings, Microsoft announcing plans to triple data center capacity within six years, Nvidia once again raising its market size estimate, and SpaceX eyeing over $100 billion in sales this year with plans for an orbital data center in 2027. "The giants are accelerating off a much bigger base." Nvidia currently trades at just 10.3 times its projected 2028 earnings. Wilson's ideal entry point: yields stop climbing, the calendar flips, and the market gets room to look toward 2028, at which point the leaders can reignite.
Direction Two: Commodities - Everyone Debates Currencies, No One Makes the First-Order Trade. Given the current environment marked by high nominal growth, deteriorating fiscal outlooks, currency devaluation concerns, and commodity-driven inflation pressure, Wilson says he's "surprised" at the lack of attention on commodities and related equities. "I have limited interest in energy, but plenty of conviction in metals," he comments, specifically citing copper: "Copper trades brilliantly, while its related equities haven't seen a valuation re-rating despite scarce growth capex and increasingly visible supply tightness." His call is blunt: pundits write sprawling essays about the dollar's collapse, but the copper pit is where the actual trade is.
Direction Three: The German Market - The Market You Think You Hate Isn't What You Think It Is. The stereotype of Germany is one of decaying factories with autos and chemicals being eaten alive by China. Wilson dismantles this with data: autos and chemicals account for only 5% and 7% of the German equity index, respectively, and domestic business makes up just 20% of DAX sales. Meanwhile, Germany is pushing massive fiscal stimulus, and Goldman Sachs projects 17% earnings growth for German corporates by 2027. The DAX composition has shifted toward defense, electrification, AI, and tech, all while trading at only 15 times earnings. Wilson flags a key catalyst: "The biggest sellers of German stocks are German savers, and pension reform is about to turn them into buyers."
Direction Four: Large Banks - Finally a Volume Story That Doesn't Depend Only on Rates. Wilson describes Goldman Sachs' Richard Ramsden's research on US banks as a "must-read." The core numbers: capital markets fee income declined roughly 1% annually from 2010 to 2019, but has compounded at 11% per year from 2019 through 2026. Even stripping out the peak impact of AI-related business, Ramsden calculates a sustainable 4% annual growth rate. Wilson also notes that AI-related financing needs could pull deal flow back from boutique shops to large banks with balance sheet strength. "Fifteen years of heavy regulation is now receding - unless the Democrats sweep in November," he writes. The central debate remains whether such high returns on equity can actually last.
Comments