Been digging into the biggest AI companies that have dominated the market, and it's wild how concentrated the landscape has become. If you're thinking about AI exposure in your portfolio, the story really comes down to understanding which players are actually building the infrastructure versus just slapping AI onto existing products.

Let's start with the obvious ones. NVIDIA basically owns the hardware layer - we're talking a company that hit nearly 4.6 trillion in market cap by designing the chips everyone needs to train AI models. Their partnership with OpenAI for that massive 100 billion dollar data center buildout is probably the most significant infrastructure play we've seen. When you're funding someone else's superintelligence ambitions, you're betting on a future that hasn't even been built yet. That's the kind of conviction that defines the biggest AI companies right now.

Microsoft's approach is different. They've gone all-in on OpenAI integration - not just as an investor but by building Copilot and embedding it everywhere from Windows to their entire 365 suite. The 80 billion they committed to AI infrastructure this year shows they're not messing around. Azure's been the real beneficiary here. Their market cap staying above 3.9 trillion even through the March volatility tells you the market believes this bet is paying off.

Alphabet's in a similar position but with more optionality. They've got Gemini, custom chips, partnerships with NVIDIA on robotics and drug discovery. Hitting nearly 3 trillion market cap, they're positioned as one of the biggest AI companies but they've got more diversification than pure-play AI bets. That's both a strength and maybe a slight weakness if AI becomes the only thing that matters.

Now here's where it gets interesting - the Canadian side. CGI and OpenText are flying under the radar for a lot of US-focused investors. CGI's doing serious work in enterprise AI integration, especially through their Google Cloud partnerships. OpenText just sold off their eDocs unit specifically to focus on AI capabilities. These aren't the flashiest names, but they're among the biggest AI companies in their markets and actually shipping real solutions to real customers.

Propel Holdings is the wildcard - a fintech leveraging AI for lending decisions outside traditional banking. The fact they ranked sixth on the TSX30 with 247 percent revenue growth says something about where the real business value is being created. It's not always the trillion-dollar names.

Down under, you've got NextDC with that NVIDIA DGX certification, Megaport building the network layer for AI infrastructure, and Nuix doing something genuinely interesting with deep learning for document classification. Australia's AI ecosystem is way more developed than people realize.

The biggest AI companies aren't necessarily the ones with the highest market caps - they're the ones controlling chokepoints. NVIDIA controls the chips. Microsoft controls enterprise integration. Alphabet controls search and consumer products. Everyone else is either building specialized solutions or infrastructure supporting these giants. If you're building an AI portfolio, you need exposure to both the picks-and-shovels players and the ones actually deploying AI at scale. The market's still figuring out which companies will actually capture value from this shift, and that's what makes it interesting right now.