The first half of the year broke most of the old market rules. "Sell in May and go away" failed. Stocks kept climbing through a global repricing of oil and a global repricing of rate expectations, even with war in the Middle East. Inflation expectations, measured by break-evens, fell hard. The labor market stayed strong despite AI fears.
The big lesson: disruption
The one thing to carry into the second half is disruption. AI is reshaping the market, and the tech sector shows it most clearly. Leadership has changed hands in a way many investors miss. The Magnificent 7 is still negative on the year. Semiconductors just posted their best quarter on record. Software inside tech had a terrible Q1 and has barely recovered, still down about 15%. The takeaway: watch the trends under the hood, because the usual playbook keeps getting challenged.
Small caps: quiet strength
The Russell 2000 had its best first half since 1991. What matters more for small caps right now, hyperscaler capex spending or the Fed? Neither is the real story. Past small-cap rallies came in fits and starts and did not last. This time these companies can handle rates where they are, and earnings there have been very good. The small-cap space is up over 20%, beating the NASDAQ.
The channel most people overlook is margins. Small caps run operating margins roughly half those of large caps. AI productivity flows straight through the margin line into earnings, so the lower-margin small-cap group has the most room to gain. With earnings this strong, the current level of rates is not a problem, which is why the performance has held up even through recent rate swings.
Why do investors overlook this? Size, mainly. The S&P 600 holds about $2 trillion across all 600 companies. Dropped into the S&P 500, it would carry roughly a 2.5% weight, about the same size as Broadcom, so it rarely stays top of mind. Its sector mix also differs from the mega-cap tech world: more industrials and financials, spread across smaller weights. That gives real diversification away from large-cap tech. Small caps are less followed and barely represented in passive index portfolios, so their strong run in an environment where rates have not come down surprises people.
The "lag seven" and a possible rotation
Multiple compression has hit the biggest names, now sometimes called the "lag seven." Many now argue Nvidia looks cheap, with one claim that its P/E is lower than Hershey's. Where is the catalyst to rotate back in?
Large-cap P/E ratios are down 10% over the year even with index values high. The MAG 7 specifically has seen P/E ratios drop 20%. The rotation may have already started this quarter, showing up as a clear split: when semiconductor and memory names sell off, the MAG 7 rises, a bifurcated trade.
This matters heading into earnings season, starting next week. Last quarter delivered some of the biggest surprises in recent memory, around 17-18% for the S&P 500, and earnings growth is already expected to be high. Investors will be more sensitive than usual and will lean toward the more discounted parts of the market. The MAG 7 now fits that category, which was not true for years. Expect buyers to nibble at beaten-down areas, the MAG 7, software, and other tech names left out of the AI infrastructure rally, on any profit-taking in the momentum trade.
Sorting good software from bad
Software underperformed on the year yet carries high earnings expectations, and some names were arguably swept up wrongly in the SaaS selloff. What separates a software name worth owning from one to avoid?
Look for deep workflow integration that a better technology cannot simply undercut, plus very high switching costs. AI will help many software firms push their products to a wider audience and turn that into real productivity gains and higher operating margins, the same margin channel that drives earnings. The winners are the ones with installed customer bases that can fold AI into their software rather than be replaced by it. That takes deep, active management to find. All of these companies cannot be worth what the weighted-average performance of the space implies, and given how discounted valuations are, this is a good place to hunt for opportunities.
