From Barbell to Conviction, and Back Again
At the start of 2026, the most sensible posture for an investor was caution expressed through balance. The year was set up to be far more volatile than the prior one, and a barbell portfolio — meaningful exposure to value stocks on one side and to technology and artificial intelligence on the other — was the right way to hold conviction while still being able to play that volatility. The point of the barbell was never to abandon either category. It was to always own some value and some growth, but to lean into whichever side the swings of the market made cheap.
By late March, the market handed investors exactly that opportunity. The technology and AI complex had sold off hard, and the discount became too large to ignore. Growth was trading at a 20% to 25% discount to fair value — a margin of safety wide enough to justify taking profits in value and reallocating aggressively into growth, technology, and AI specifically. That call proved correct, and then some. Growth stocks subsequently exploded higher, rising more than 20%, while the broad U.S. technology index climbed more than 32%.
That is precisely why the discipline now points in the opposite direction. The margin of safety that justified the overweight has largely closed. Value stocks today screen as roughly 7% undervalued — slightly cheaper than they were back in March — while growth is only about 5% undervalued, a fraction of the discount that existed at the lows. This is not market timing or trading; it is long-term positioning driven by where valuations actually warrant capital. When the gap collapses, the overweight in technology no longer makes sense, and the prudent move is to take money out of tech and AI and rotate it back into value.
The Critical Distinction Within AI
The most important nuance for any investor today is that "AI" is not one trade. There is a sharp dislocation in valuations within the theme, and conflating the two halves of it is dangerous.
The companies at the genuine technological forefront of AI still have a long growth runway. Names like the leading accelerator and networking franchises — and the dominant search-and-cloud platform — remain significantly undervalued even after enormous runs. When you build a disciplined discounted-cash-flow model with a five-to-ten-year explicit forecast period and feed in the kind of growth those businesses are likely to deliver over the next several years, there is still plenty of room left in the share price.
The danger lies in the commodity-oriented layer: memory chips, optical and networking hardware, and similar semiconductor names that have gone gangbusters this year. Most of these now look meaningfully overvalued. The logic is the oldest in commodities: the cure for high prices is high prices. There are real shortages today, but the best fix for an overvalued, supply-constrained market is more supply — and more supply will come on over time. As it does, prices fall, margins compress, and the earnings that justify today's multiples evaporate. Even when you model huge growth straight through 2028, the models point to top lines and margins beginning to roll over in 2029 and 2030 for these hardware names. If there is any slowdown at all, those stocks could gap sharply lower. That makes commodity hardware the single clearest place to take profits right now.
The over-ordering risk amplifies this. The data center buildout is a food fight. Purchasing managers are effectively told that cost is irrelevant — just secure the hardware, because a multi-hundred-million-dollar facility will not be delayed for want of memory. That urgency inflates current demand in a way that may not be sustainable, and it raises the question of whether the buildout is even happening faster than the power infrastructure required to run it.
The Macro Backdrop: Rates and Sticky Inflation
The risks weighing on this market are largely the same ones identifiable at the start of the year, and they have not eased. Inflation is running hotter than expected and is not coming down soon, with Brent and WTI crude both well above $100 a barrel and no resolution in sight to the conflict involving Iran. That keeps the central bank firmly on the sidelines; rate cuts are off the table for the foreseeable future.
Interest rates themselves deserve close watching, even if they are not yet alarming. The 10-year yield sits around 4.66%, having touched roughly 4.69% intraday, and the 30-year is at its highest level since 2007. The threshold that matters is a 10-year approaching a 5% handle. At that point, long-duration, liability-matched investors — pension funds and insurance companies — would have real incentive to reallocate out of equities and into bonds. That rotation would, in turn, drain valuation support from growth stocks, whose long-duration cash flows are the most sensitive to higher discount rates.
A second, less-discussed pressure point sits offshore: Japanese government bonds. The Japanese 10-year is at its highest yield since 1997. Currency intervention by Japanese authorities appears short-lived — the yen is weakening again, and a move toward 160 may force renewed intervention — and it has not been enough to keep the bond market in check. Given Japan's debt-to-GDP burden, this is a genuine source of concern for global markets.
Where Value Actually Lives
Rotating into value is only useful if there are concrete places to put the money. Several stand out.
Financial services is becoming more attractive. The most undervalued large bank — though not necessarily the highest-quality one — offers compelling value at current prices, and an independent advisory platform with strong growth dynamics is among the best picks in the sector.
Real estate remains one of the more attractive corners of the market, with important caveats. East Coast urban office space, particularly in New York, faces a difficult path to refilling vacant floors and is best avoided. Urban office in Texas and Florida looks comparatively healthier. The safer way to own the sector, however, is through defensively positioned names — a diversified net-lease REIT and a healthcare-focused REIT both look undervalued today.
Consumer staples require careful separation of perception from price. The big-box defensive retailers have had great runs and benefit genuinely from a backdrop in which low- and middle-income households are under spending pressure and trading down from traditional grocers, lifting foot traffic, same-store sales, and margins. But valuation is the problem: the dominant discount retailer trades around 45 times this year's earnings estimate, and the warehouse club well over 50 times — richer than many of the AI names investors fret about. Both carry the lowest risk-adjusted rating, and both are worth steering clear of. The opportunity in defensives sits instead in food names, many of which look very cheap as the GLP-1 weight-loss-drug headwind appears to be fading. A leading global snack maker is a standout, with the best exposure to fast-growing emerging markets — roughly 45% of its top line — making it attractive on both valuation and dividend income.
The Takeaway
The throughline is discipline over momentum. A parabolic run in growth has erased the margin of safety that once made technology the obvious overweight, so the rational response is to take profits — especially in commodity hardware — and rotate into value, financials, defensive real estate, and cheap food names. The genuine AI leaders can still be held for their long growth trajectory, but everything around them should be judged against the same unsentimental question: does the long-term model still justify the price? Where it does not, the time to act is before the supply arrives, the rates rise, or the over-ordering reverses — not after.