Diversification pays again in 2026
Diversification is the last free lunch open to all investors. For long stretches, sometimes decades, it fails to pay off, and simply owning the S&P 500 or the Mag 7 was the winning bet. In 2026 that flips. The real payoff now comes from spreading money out: overseas markets, smaller companies (down in capitalization), and value stocks.
This part of the market cycle is where growth at a reasonable price should lead. The Mag 7 has taken an odd spot this year - the back of the bus. Strong company earnings and inflation trends push money into more diversified portfolios, and this trend should hold for years.
AI is real, but look past the pure plays
AI is real and you must own it. Also buy companies that gain from AI - the second and third-order beneficiaries. As AI skepticism grows and the boost behind pure-AI names fades a little, investors will shift toward these secondary winners. That broadening is the power of diversification at work.
Buy both the main hyperscalers and the names that gain further down the chain.
Earnings were mixed, and caution is real
Tech earnings, now mostly reported, were mixed. There were bright spots. CapEx spending looks fine as long as it earns a return, and some companies showed that better than others. Frustration and caution remain. The roughly $800 billion in spending that everyone knows about draws skepticism on both return on investment and the timing of when payoffs arrive.
Growth stocks do not have to stop growing to fall. If a growth line runs at a 45-degree angle and then eases to 35 degrees, the price drops to match. When growth slows, sell-offs hit the stock.
Why "air pockets" form in growth names
Growth managers own these names. When growth slows, a value manager does not step in to buy the stock at those prices. The growth manager must sell to another growth manager who still believes in the name. With no value buyer present, air pockets open up - sudden price gaps that are hard for investors to stomach. That mechanism drove much of the summer volatility in these names.
Are Mag 7 names value plays?
Some are, depending on earnings and history. Nvidia (NVDA) reports after tonight's close and had fallen in seven of the past eight days, with many investors cautious about the report; calling it a value stock is a stretch for now. Apple (AAPL) clearly became one - it drew Warren Buffett with its long-term vision, its hold on customers, and what is the most coveted real estate on the planet, the palm of the hand. It sold down to levels that attracted value managers. Apple is an incredibly mature company, so such cases are rare. Most other AI stocks are nowhere near value - still growth names, some still growing, not yet profitable.
Where the second-order winners hide
The beneficiaries span many areas: companies with high employee counts, industrials, anything where AI can take over work. A key measure going forward will be revenue per employee. Target companies with the potential for very high revenue per employee, which lets them use AI to lift profit margins. This is less about picking sectors, since each sector will have adopters and non-adopters, and more about finding firms that keep adopting AI regardless of sector.
There are two sides to the opportunity: building AI out (infrastructure, utilities) and using it (specialized software, companies that apply AI). Both offer openings, though deciding what will work is hard, and the market itself is still trying to sort it out.
This is not 1999
This market is nowhere near the tech bubble. The PE multiple on the NASDAQ 100 in 1999 was 100 times - triple digits. Back then euphoria and optimism ran wild, and people would pay any amount for anyone who even owned a domain name, because the internet was the next thing. No one called it a bubble at the time.
The skepticism inside AI today is healthy, and the rotation and corrections are good. Markets need time to digest and question big moves. When people openly debate whether something is a bubble, that debate itself argues against a bubble. This is not that.
