Just been thinking about how you don't actually need a ton of capital to build a solid artificial intelligence portfolio. Like, $2,000 is honestly enough to grab some real positions in companies that are genuinely pushing AI forward.

I've got three main holdings that I keep coming back to. First is Nebius — probably the least talked about of the bunch, but honestly it might have the most upside. They're this Dutch company building out AI cloud infrastructure that developers and the big hyperscalers actually need. Their annualized revenue hit $1.25 billion last year and they're projecting $7-9 billion for this year. That's the kind of growth trajectory that gets my attention. They just picked up Tavily too, which is an AI search platform, so they're expanding what they can offer to enterprise clients.

Then there's Nvidia. Yeah, everyone knows Nvidia at this point. Market cap is sitting at $4.6 trillion but the thing is, they're still growing like crazy. Last quarter showed 62% revenue growth hitting $57 billion — and most of that, like $51.2 billion, came straight from data center sales. When you've got Microsoft, Alphabet, Amazon, and Meta collectively planning to drop $650 billion on AI infrastructure this year alone, Nvidia's just going to keep printing money.

The third one is Palantir. Different angle here — they're not building the chips or infrastructure, but their AI software is legitimately some of the best out there. Their platform pulls from hundreds of data sources and gives you real-time insights. The AIP (their AI Platform) is the real driver, and it's working. They did $4.475 billion in revenue last year (up 56%) and they're guiding to $7.2 billion this year. That's 60% growth.

If I was allocating $2,000 across these three, I'd probably weight it something like: Nebius at 25%, Nvidia at 50% (it's the safest bet), and Palantir at 25%. These are the companies actually building the infrastructure and software that AI needs. Worth looking into if you're trying to get exposure to artificial intelligence without overthinking it.