Just looked at the Nvidia bull case again and honestly the upside still feels underpriced. Most Wall Street analysts are sitting at a $265 target, which is like 47% from here. But if you've been paying attention to how much hyperscalers are actually willing to spend on AI infrastructure, you know consensus always gets it wrong.

Here's what actually matters: Nvidia doesn't just make the best chips. That's the surface level take. What really gives them the moat is the full stack - they control the hardware, the networking, and they've spent 20 years building CUDA into this ecosystem that makes it incredibly expensive for anyone to switch. Developers are locked in, companies are locked in, and the total cost of ownership actually favors Nvidia even though the upfront price is brutal.

The financial picture is pretty clean too. Q4 showed 73% revenue growth and they're guiding for acceleration again. Gross margins expanding. Earnings up 82%. And the addressable market isn't even close to saturated - data center GPU spending is supposed to grow 35% annually through 2033. Beyond that you've got autonomous vehicles, robotics, every major OEM using Nvidia tech.

Jensen Huang has been pretty explicit about where this is headed - he's talking $3-4 trillion annually in AI infrastructure spending by 2030. That's the scale we're talking about.

If Nvidia just maintains its current valuation multiple and hits earnings estimates, you're looking at $285 by end of year. That's 58% upside from $180. Honestly I think that's the floor, not the ceiling, but even the conservative case looks pretty compelling for the next 7 months.

The way I see it, Nvidia is still the only real way to play the infrastructure buildout. Everyone's trying to find the next thing but the moat here is just too wide.