Nvidia: A Perfectly Reasonable Prediction

by

My aunt Carol, who invests solely in companies she sees advertised on daytime television, recently cornered me at a family barbecue. She wanted to know if I thought Nvidia was “a good thing.” I explained, as patiently as one explains anything to someone who still believes in the healing power of crystals, that it makes the little brains inside computers think faster. She nodded, then asked if it made a decent toaster oven. It was at that moment, watching a bee circle her floral print dress, that I realized the market has no idea what’s coming.

Everyone’s fixated on the hype, the AI gold rush. I prefer to think of it as a very expensive game of digital solitaire, and Nvidia, for now, is holding all the aces. They’re predicting growth, of course. Everyone predicts growth. It’s practically a contractual obligation for any publicly traded company. But I’m looking at 2026, and I suspect they’ll surpass the consensus revenue estimates of $323.3 billion. Not because of some revolutionary breakthrough, but because the alternative – everyone suddenly deciding they don’t need increasingly complex algorithms to suggest things they already want to buy – seems…unlikely.

Loading widget...

They’re moving beyond just selling chips, you see. It’s the whole package now – the “rack-scale solutions” as they call it. Which sounds suspiciously like they’re just bundling everything together and charging a premium. Like a hotel charging extra for the mini-bar. But it works. They’ve got these hyperscalers – enormous data centers filled with blinking lights and the quiet desperation of overworked IT professionals – practically signing contracts in advance. They’re saying they have over $500 billion in revenue visibility through 2026. That’s…a lot of blinking lights. And a lot of pressure on those IT professionals.

And the shift to “inference workloads” – which, as far as I can tell, means making the AI actually do things, instead of just thinking about doing things – is going to drive another wave of upgrades. They’re promising lower total costs of ownership with their new Rubin systems. That’s what they always promise. But people are suckers for efficiency. Especially when someone else is paying the bill. I remember my father replacing perfectly good appliances simply because the new models had more buttons. It’s a family trait, apparently.

So, yes, I think they’ll beat expectations. Not because of genius or innovation, but because human nature is remarkably predictable. And because my aunt Carol, bless her heart, is probably already buying stock.

Margins: A Delicate Balancing Act

They’re also claiming they can maintain gross margins around 75%, despite the competition. That’s ambitious. It’s like trying to keep a soufflé from collapsing. But they’re selling high-margin data center GPUs and networking products. And software. Everyone’s selling software now. It’s the digital equivalent of adding a surcharge for breathing. Advanced Micro Devices and Qualcomm are trying to muscle in, of course. But Nvidia has a head start. And a certain…momentum. Like a runaway shopping cart filled with expensive technology.

Market Share: Holding the Line

They lost a couple of percentage points in GPU market share last quarter. A mere two points. The analysts are making a fuss. It’s like complaining that your artisanal coffee isn’t quite hot enough. But they still control 92% of the market. That’s…significant. They’ve built this ecosystem around their CUDA software. It’s like a walled garden. Beautiful, meticulously maintained, and difficult to escape. Switching to a competitor isn’t just about the hardware. It’s about retraining your staff, rewriting your code, and admitting you made a mistake. And no one wants to do that.

So, while the competition will undoubtedly intensify, I suspect Nvidia will manage to hold onto its dominance. Not because they’re the best, but because changing is hard. And people, generally, prefer comfort to progress. It’s a flaw, I suppose. But it’s also a remarkably reliable investment strategy.

Read More

2026-02-03 07:32