Citi's Tyler Radke Says AI Is Splitting Software in Two. The Braze Deal Backs Him Up.
Citi's co-head of US software equity research says AI is separating software's haves from its have-nots, with Microsoft growth topping 20% while Adobe absorbs budget pressure. Salesforce's agreement to buy Braze at $47.50 a share is the clearest sign yet of where the consolidation is heading.
Citi's Tyler Radke thinks AI is splitting software into winners and losers, and the market is already pricing that split in. He made the case on Bloomberg Television just as Salesforce agreed to buy Braze for $47.50 a share.
That pairing isn't a coincidence. It's the whole story in miniature.
How We Got Here
Rewind a few months. According to Radke, who co-heads US software equity research at Citi, the sector looked very different earlier this year. His stat is brutal: the worst-performing chip stock in his coverage universe still beat the best-performing software name. Not narrowly. The laggard beat the leader.
That's an inversion, and it played out through the spring and summer as money rotated into anything touching compute and power and away from anything selling seats. Software multiples compressed. The narrative moved from "recurring revenue is king" to "who's actually getting paid for AI."
Then the deal flow showed up. Salesforce agreed to buy Braze at $47.50 a share, which is the kind of transaction that happens when a strategic buyer decides organic growth alone won't get it where it needs to go.
The Split, Explained
Radke's numbers sketch both sides of the divide. He expects Microsoft's revenue growth to top 20%. Twenty percent, at that scale, is a figure most software CEOs would take in a heartbeat.
On the other side sits Adobe, which he says feels pressure as AI projects absorb corporate budgets. Here's the mechanism, and it's simpler than most of the commentary suggests. Software budgets aren't infinite. A CIO writing checks for AI pilots and compute capacity has less left over for seat expansions and upgrades at the vendors that used to get the first call.
So the money isn't vanishing. It's moving.
Plenty of the AI winners-and-losers chatter treats this as a permanent structural shift. Color me skeptical, but I think a real chunk of it's a budget timing story dressed up as something bigger. Granted, the timing itself is doing genuine damage to revenue forecasts. Admittedly, that's cold comfort if you own the stock.
The question worth asking: is AI expanding the software pie, or just sliding the slices around? Radke's framing implies the second, at least for now.
What to Watch
Microsoft's next earnings report lands in late January. That 20% growth number is the cleanest available test of the thesis, because Microsoft carries both AI infrastructure revenue and traditional software seats under one roof. If the seat businesses hold while AI climbs, the pie is growing. If seats wobble, the reallocation story wins.
Adobe's next print matters just as much, from the opposite direction. Watch net new annual recurring revenue and any commentary on enterprise budget cycles. Two soft reports in a row and the skeptics get a lot louder.
And keep an eye on the deal tape. Salesforce just showed what it'll pay for a customer engagement platform, $47.50 a share, and there's no shortage of mid-cap software names trading below where they started the year. Deals get done when the gap between public valuation and strategic value gets wide enough, and that gap looks wide right now.
The split Radke describes isn't a verdict on AI. It's a verdict on which software businesses actually get paid for it. Right now, the market is voting with real money.
