So I've been looking at where to park $5,000 if you've got some dry powder right now, and honestly the AI infrastructure play is still looking pretty solid. Let me walk through what I'm seeing in the market.

First, the obvious one - Nvidia. Yeah, I know everyone talks about it, but there's a reason. They've basically become the picks and shovels supplier for the entire AI gold rush. Their GPUs are still the primary computing units for training and running AI models, and that demand just isn't slowing down. Wall Street is projecting 52% growth for fiscal 2027, which just kicked off. Some people worry we're in an AI bubble, but realistically, even if the hype eventually cools, Nvidia's positioned to do fine. They're not betting on AI being revolutionary - they're profiting from the infrastructure build-out regardless.

But here's where it gets interesting. Broadcom is making a real play to chip away at Nvidia's dominance. They're not trying to out-GPU Nvidia - instead they're building ASICs, these application-specific chips optimized for AI workloads. The performance difference can actually be better than GPUs for certain tasks, and they're cheaper to produce. For Q1, Broadcom's expecting their AI semiconductor revenue to literally double year over year. That's faster growth than Nvidia's seeing right now. Will these chips completely replace GPUs? Probably not. But they're definitely carving out market share, and there's room for both companies to win here.

Then there's Taiwan Semiconductor Manufacturing. TSMC is basically the foundry that manufactures all these chips that Nvidia and Broadcom design. They've got the tech and capacity nobody else can match. If you want a more neutral play on the AI infrastructure build-out without picking sides between chip designers, TSMC is the move. They win as long as AI spending stays elevated, and most projections have that running through at least 2030. Wall Street's modeling 31% growth this year and 22% next year, so the numbers are still attractive.

Now Microsoft is interesting for a different reason. They're playing both sides - they've got Azure growing at 39% year over year as of their recent quarter, and they're sitting on a $625 billion backlog in that business unit. That's serious runway. The stock took a hit recently though, which actually created an opportunity. It's trading at 25 times forward earnings, which is the lowest valuation we've seen in a while. That's the kind of dip where you want to be a buyer.

The common thread here is that all these companies are positioned to benefit from sustained AI spending over the next several years. If you've got $5,000 to invest and you're not going to need it for bills or emergencies, these look like solid plays that should outperform the broader market. The best stocks to buy right now are the ones positioned at the infrastructure layer of this AI build-out, and these four companies represent different angles on that same thesis. Worth considering if you're looking at where to deploy capital.