Nvidia’s Shadow: A Few Diversions

by

Nvidia. It’s the kind of stock my Uncle Phil would have mortgaged the house for, then loudly complained about at Thanksgiving. It’s still going strong, of course. Everything’s “disruptive” these days, and Nvidia, for the moment, is the disruptor’s disruptor. But there’s a certain…pressure, watching a stock climb like that. It reminds me of trying to hold onto a beach ball in a hurricane. Eventually, something’s going to give. Not that I’m predicting a crash, just suggesting a little…leisurely drifting toward other options. Diversification, they call it. I call it admitting you don’t have the stomach for all this.

I’m not saying abandon ship, mind you. More like…pack a smaller bag. Consider a side trip. Alphabet and Broadcom, for instance. They’re not quite as flashy, but they offer a certain…stability. Like sensible shoes for a very volatile dance floor.

1. Alphabet

Everyone thinks of Google as the search engine, which, let’s be honest, is still pretty good at finding pictures of cats. But they’re doing a lot more. They’ve got this thing called Gemini, a large language model. Sounds intimidating, doesn’t it? Like something out of a science fiction novel. But it basically means they’re trying to teach a computer to write like a human. Which, after reading some of the comments sections online, seems like a remarkably low bar.

Loading widget...

They’ve also designed these Tensor Processing Units, or TPUs. Apparently, they’re cheaper than Nvidia’s GPUs. Which is good, because everything is expensive these days. My niece asked me for $20 for a “virtual pet” the other day. A virtual pet! I tried explaining that we used to entertain ourselves with sticks and rocks, but she just looked at me like I was a relic from the Paleolithic era. Anyway, these TPUs are custom-designed for specific tasks, which is a fancy way of saying they’re good at doing one thing really well. Like my Aunt Mildred, who can perfectly identify any species of bird, but can’t remember where she parked the car.

Alphabet’s revenue increased 18% last quarter, and cloud services are up 48%. Which is impressive, I suppose. I’m more impressed when I manage to find matching socks. Still, it suggests they’re doing something right. Or at least, something that’s generating a lot of money.

2. Broadcom

Broadcom started out making networking stuff. I don’t pretend to understand what that entails, but it involved a lot of wires and blinking lights. Now they’re getting into AI, which is apparently the future. Or at least, the future until something even newer comes along. They’ve developed their own application-specific integrated circuits, or ASICs. Another acronym. It’s exhausting. These ASICs are specialized chips that handle specific tasks. See a pattern here? Everyone’s making chips that do one thing really well. It’s like a world populated by savants.

Loading widget...

Their revenue increased 28% last quarter, with AI semiconductor revenue up 74%. They also generated $26.9 billion in free cash flow. Which is a lot of money. Enough to buy a small island, probably. Or at least, a very nice condo. They’re also profitable, which is always a good sign. It means they’re not just burning through cash like a teenager with a credit card.

Broadcom has robust non-AI businesses, like broadband. Which is reassuring. It means they’re not entirely reliant on the whims of the AI market. It’s like having a backup plan. Or, in my case, a second pair of glasses. It’s comforting to know you’re prepared for anything. Or at least, for the inevitable failure of technology.

Read More

2026-03-19 05:02