Skip to main content

Myainote

 · 

You Think You're Buying Robots with NVDA? 92% Is Data Center; Robots Are a Rounding Error

Every time the robot theme heats up, someone says: just buy NVIDIA, it's "the brain of robots." Half true, half illusion. Latest quarter: ~$81.6B revenue, ~92% data center; robot-related is a low-single-digit rounding error. Buy NVDA "for robots" and you own a bet on data-center capex. Worse: if you hold mega-cap tech, an AI ETF, or semis, you're already long it — buying more just levers the same AI-capex + rate-duration factor.

Every time the robot theme heats up, someone tells me: play it safe, just buy NVIDIA — it’s “the brain of robots,” it wins no matter who wins.

Half true, half illusion. Look at the latest quarter first:

NVIDIA’s total revenue was ~$81.6 billion, of which data center was ~92%. Robot-related revenue sits inside the “edge / robotics” segment — which is itself only a low single-digit share of the total, and robots are a small slice of that. Net it out, and robots contribute a low-single-digit-percent rounding error to NVIDIA’s current revenue.

So when you buy NVIDIA “for robots,” what you actually own is a bet on data-center capex. Its stock is driven by the AI capex guidance of Microsoft, Google, Amazon, and Meta — not by humanoid progress.

Robots are NVIDIA’s “free option” — but don’t pay a cent extra for it

To be clear: NVIDIA genuinely sits on the compute/simulation upstream of robots, and long-term it’s a beneficiary. Robots are a free, un-priced option for it — if humanoids take off, it catches the upside without you paying for it.

That “without paying” is the whole point. Buying more NVIDIA for the robot narrative means paying a premium for a free option while taking on data-center capex-cycle risk — two different things.

The more hidden trap: you may already be overweight it

Here’s the account most people skip: if your portfolio already holds mega-cap tech, an AI ETF, or semis — you’re almost certainly already heavily long NVIDIA, indirectly.

Buy another slug “for robots” now, and you’re not diversifying into a new theme — you’re levering up the AI macro bet you already hold. Nominally you added a “robot name”; substantively you just deepened the same “AI-capex + rate-duration” factor exposure. Real diversification, it does not give you.

So before buying NVIDIA, do two things: one, compute your look-through exposure — add up all the NVIDIA hiding in your funds and ETFs, and see how much you already own; two, admit honestly that you’re buying a data center, not robots.

Great company. But “great company” and “should I add for this narrative” are two separate questions. Fame bundles unrelated things and sells them to you — your job is to take them apart.


Have you ever computed your “look-through NVIDIA exposure”? Add the funds and ETFs — it may be more than you think.

(Independent industry and educational research, not investment advice. Companies named and sell-side estimates are illustrative examples, not recommendations. Data is from public sources and may change. The author and affiliates may hold positions in securities mentioned.)

— Adapted from Embodied Intelligence Investing, Deep Dive IX: NVDA Factor Overlap & SYM Concentration

#NVIDIA #EmbodiedAI #AInvesting

Download Pickful App

Better experience on mobile

iOS

Android

APK