UBS Cut the Big 6 and Loved 14 Tech Names — Only One Giant Made the List

Try a gut check on what "tech" means now, and you land on a number that would have startled a 1999 investor. The ten largest companies in the S&P 500 make up almost 40% of the index's value — the most concentrated this market has been since the dot-com bubble, when the top ten carried roughly a quarter and, back in 1980, barely 15%. Every one of today's top ten is worth at least a trillion dollars.

Now set that next to what UBS did in the week of September 18. The bank cut its sector recommendation on the "Big 6" — Alphabet, Amazon, Apple, Microsoft, Nvidia, Meta — the six megacaps whose sheer size is the reason for that 40% number. Then it published a list of 14 technology, media, and telecommunications stocks it wants to own.

The two signals read as if they cancel each other out. They answer different questions. The sector cut is a market-cap question: when six stocks are that large, "tech" as a category is really a one-way bet on them, and their size is the problem. The stock list is a where-is-the-money-being-spent question. UBS, like the rest of the Street, thinks the money is going into AI infrastructure, and it thinks the cycle is early — only 8% of companies it surveyed have deployed "agentic" AI at scale, and it models data-center capital spending growing at least 30% a year for the next three.

Here is the tell: only one of the 14 picks, Amazon, is a Big 6 name. The rest are a scavenger hunt around the same AI buildout that avoids the index's center of gravity. The list splits roughly into people who build the hardware and people who sell the software.

On the hardware side, UBS's top pick is Texas Instruments — the plain-vanilla analog chip maker, up 3.29% the day the note landed. It sees a rising demand cycle and revenue growth near 20%. Celestica builds the AI servers themselves and is modeled for 70% growth. TD Synnex, through its Hyve Solutions unit, supplies the custom servers and networking gear.

On the software side sits Palantir, which UBS calls "very attractive" at $177.64 against a $250 target, growing revenue 88% with margins in the low 60s, with JFrog and Procore alongside. The remaining names fit neither bucket neatly — Mastercard, Global-e, Spotify, Pinterest, AT&T, EchoStar, SS&C. Several of those, Barron's noted, are exactly the tech "laggards" UBS is betting will catch up.

The valuation spread does the real admitting. Amazon is the cheap one — $253.71 as of this writing, a $2.7 trillion market cap, about 39 times forward earnings, and a $318 target implies the 25% upside UBS ties to AWS. Palantir is the expensive one — roughly 395 times forward earnings, about 70 times sales, a $427 billion market cap on a stock that's flat for the year even after a 120-day bounce of 24%.

Read the two signals in sequence, not together. The warning is about size — a top ten carrying 40% of the index at a concentration unseen since the bubble. The wish list is about the buildout happening underneath it, in names that mostly don't move that index at all. The punchline, per Barron's: three of UBS's favorites — Palantir, AT&T, Spotify — are also the tech laggards UBS is betting will catch up. The firm is cold on the very stocks that broke the index, and warm on a list where only one of them shows up. Go figure.