Jeff Brown has been watching Tesla longer than most. He called it before the market re-rated it from auto to AI. He visited SpaceX Starbase in south Texas. He tracks the supply chain because he spent two decades inside it — at Qualcomm, Juniper Networks, and NXP Semiconductors.

The 70X AI Agent promo is the latest expression of a thesis he has been building for years: that the biggest technology transitions are never visible at the brand level. They happen two layers down in the supply chain, where a single component manufacturer becomes the bottleneck for an entire industry.

Brown’s argument is that Tesla’s Full Self-Driving system is a general-purpose physical AI agent that happens to have been trained on driving first. The same neural network architecture that navigates a Tesla through San Francisco traffic powers an Optimus robot navigating a factory floor. The compute requirements are identical: real-time perception, decision-making, and physical actuation processed in milliseconds.

The investment angle is the company that makes the component both systems depend on — the optical hardware that gives FSD and Optimus their eyesight.

The FSD-as-AI-agent reframing

Most people think of Tesla’s FSD as a self-driving car feature. Brown thinks of it as a platform. The neural network has been trained on billions of miles of real-world driving data. It processes visual input, makes decisions, and executes physical actions. Those three capabilities — perception, decision, actuation — are the definition of a general-purpose AI agent.

The difference between FSD and something like ChatGPT is that FSD operates in the physical world. ChatGPT manipulates text. FSD manipulates a two-ton vehicle moving at highway speeds. The stakes are higher, the latency requirements are tighter, and the hardware requirements are more demanding.

Brown’s thesis is that the same platform extends naturally to Optimus. The robot is FSD’s neural network in a different body. Same vision system, same decision-making architecture, same physical actuation requirements. The bottleneck is the hardware that lets the AI see and interact with the physical world.

The hidden supplier structure

Brown has used this structure before. In 2016, he called Nvidia an AI company when Wall Street saw it as a gaming GPU company. The same hidden-supplier structure drives his Brownstone Research track record, which traces the longer arc of his supply-chain calls. He understood the CUDA architecture’s potential for machine learning before the market did. That call returned roughly 25,000 percent from his entry point.

The structure is the same here: find the component that makes the headline technology possible, identify the company that makes it, and invest before the market connects the dots. The difference is that in 2016, the hidden supplier was Nvidia itself — a $30 stock that the market had misclassified. Today, the hidden supplier is the one making the optical hardware that FSD and Optimus both depend on. Past performance does not guarantee future results. The returns cited in this article are calculated on public market data from publicly stated entry points, and the exact figures depend on the entry and exit points used.

Tesla’s FSD system uses a camera-based vision architecture. Eight cameras around the vehicle provide 360-degree visibility. The neural network processes the visual data in real time. The system does not use lidar or radar for the primary perception stack — it is vision-only, which means the quality of the camera hardware determines the quality of the perception.

The same applies to Optimus. The robot needs to see and navigate its environment. The vision system that works for a car on a road works for a robot in a factory. The same supplier, the same components, a much larger addressable market. The Tesla FSD optical supplier sibling piece breaks down the specific component layer.

The Optimus and Cybercab rollout

This is where the thesis moves from theory to observable data. Tesla’s robotaxi service is already operational in Austin, San Francisco, Houston, and Dallas. More than 700,000 paid miles have been logged. Expansion to Phoenix, Miami, Orlando, Tampa, and Las Vegas is planned for 2026.

The Cybercab — Tesla’s purpose-built autonomous vehicle — is expected to begin production. Musk has talked about building millions of them. Each one needs the same vision system, the same optical hardware.

Optimus is the bigger number. Musk has said he wants to build 1 million Optimus robots at a target price of around $25,000 each. If Tesla hits even a fraction of that target, the demand for the optical components is enormous. And the company that makes those components ships billions of units already.

The physical AI framework

The broader thesis is that AI is moving from the screen into the physical world. Software AI — chatbots, image generators, code assistants — runs on servers and manipulates digital information. Physical AI — robots, autonomous vehicles, drones — runs on hardware and manipulates physical objects.

The investment implications are different. Software AI is dominated by a small number of companies with moats in data, talent, and compute. Physical AI is a supply chain story. The companies that make the components — the sensors, the actuators, the memory chips, the vision systems — are the picks-and-shovels plays.

Brown’s background is the relevant context: he trained as an aeronautical and astronautical engineer at Purdue and spent two decades at semiconductor and networking companies, so he understands the hardware constraints better than the average analyst. When he says the optical component is the bottleneck, he is not reading a sell-side report — he has been inside the supply chain.

The 70X AI Agent promo is a snapshot of a larger transition. The idea that FSD is a general-purpose AI agent, that the same optical hardware powers both cars and robots, and that the hidden supplier is the investment — those are the pieces, and the physical AI transition behind them is real even as the question of which companies enable it remains open.

The idea is worth sitting with. Brown’s longer-arc day-one retirement plan framing pulls the same hidden-supplier logic into a portfolio built around the physical-AI transition. The 70X AI Agent trust check covers the campaign-level mechanics and refund window.