The Missing Piece for Robots

Jetson Thor is NVIDIA’s (NVDA) robotics supercomputer module, and Jensen Huang has called it “the ultimate supercomputer for the age of physical AI.” The reason it matters is easy to understand once you see the bottleneck it solves: robots have been clumsy because the compute needed to reason about physics in real time did not fit inside a robot. A robot that takes seconds to decide where to put its hand is useless. Jetson Thor packs enough on-board processing for a robot to see, understand language, plan actions, and make decisions in real time. Our full Nostradamus of AI teardown sets the stage, and the Jetson Thor and NVIDIA explainer goes deeper on the hardware.

The Four-Wave Framework

James Altucher’s Nostradamus of AI presentation uses a four-wave framework attributed directly to Huang: perception AI, which is speech and image recognition; generative AI, the ChatGPT era; agentic AI, the current phase of agents that do work for you; and physical AI, the announced next wave. It is a genuinely useful way to organize where AI goes next, and Altucher explains it better than most sell-side notes do. The “why now” argument built on it is sound: the physical world is the last domain AI has not yet colonized, and the hardware finally exists to do it. We break down that angle in our James Altucher’s Nostradamus of AI: The Pitch, Checked piece.

What the Adoption Evidence Shows

The adoption trail is concrete and checkable. Amazon runs more than a million warehouse robots, Walmart has its Symbotic deal, FedEx is testing truck loading, and BMW uses humanoids in X3 production. These are real, publicly reported programs, and they are the foundation for the pitch’s signature number, a claim that the humanoid robot market “could grow 190,597%.” Our NVIDIA physical AI piece unpacks that figure.

The honest decoding matters here. That 190,597% is the growth multiple of the humanoid robot population, from roughly 100,000 today toward an extreme upper bound of eight billion in ten years. It is not a revenue forecast or a stock return, a distinction the presentation does not make explicit, and eight billion units in a decade is far above what Morgan Stanley and other research houses publish. The number is real arithmetic on an aggressive assumption.

The Investable Names

The free pick handed over in the presentation is Marvell Technology (MRVL), and it is a legitimate one. While NVIDIA builds the AI brains, Marvell builds much of the “nervous system and plumbing” that connects them, the custom silicon and interconnect that move data between chips at speed. It is a real, liquid, large-cap way to act on the data-center side of the thesis without buying any paid report at all. The deeper picks, a tiny Optimus sensor supplier teased at “20X or more” and a SpaceX connectivity-chip partner, sit behind the $49 order form, and our robotics and NVIDIA explainer covers the supply chain that feeds the wave. The details are in our Elon Musk’s Secret AI Linchpin: The Report Examined article.

What to Watch

The honest caveats are worth naming. The authority transfer from Huang is a stretch, he makes technology predictions, not equity recommendations, and the “roadmap to riches” framing converts strategic commentary into a track record it was never meant to be. The Optimus Gen 3 catalyst is presented with more certainty than any product launch deserves. And as with all Altucher presentations, the past-winners catalog is the highlight reel, not the full record.

The Bottom Line

Jetson Thor is a real product solving a real bottleneck, and physical AI is a genuine next wave rather than a slogan. The four-wave framework is worth internalizing, and Marvell is a defensible free idea for the data-center side of it. The gap to watch is the distance between the hardware, which is real, and the specific stock returns the presentation implies, which are arithmetic on aggressive assumptions.

NewsletterVetter is an independent publication. We receive compensation from some of the services we review through affiliate links. Nothing on this site is investment advice. Always do your own research.