The biggest wrong idea investors still hold about robotics is that it means factory arms and simple car-plant machines. The real shift is broader. Tesla's (TSLA) Optimus, pushed by Elon Musk, shows the scale of what is coming. Everything runs on physical AI, which at its core means putting a high-powered computer inside a robot.
Picking up a cup of coffee shows the problem. People do it without thinking, but a robot needs thousands of physical calculations: measuring distance, measuring pressure, keeping the cup level, and not gripping too hard. All of this must happen in real time, powered by AI. The money is not only in the robots themselves. It is also in the supporting tools and parts that hold up the whole system - the "picks and shovels."
Three layers of robotics
I split robotics into three layers. The brain does the thinking. The sensors handle touch and perception, like the coffee-cup example. The testing systems make sure everything runs at top efficiency. The stakes rise with the job: a mistake in a chatbot gives you a wrong answer, but a mistake in robot-assisted surgery brings serious harm.
The brain: Nvidia (NVDA)
Nvidia (NVDA) is the obvious giant. It builds GPU technology into purpose-built computers that fit inside robots. Its big new product last year was Jetson Thor, Blackwell-chip-powered computers made to sit in physical bodies. Each handles two quadrillion operations per second. To picture that: counting up to two quadrillion, one per second, would reach back 65 million years, when dinosaurs walked the earth.
Nvidia (NVDA) also released Isaac, a robotics development and simulation platform, and Groot, a foundational AI model built specifically for robots to work out their problems inside data centers. With growing earnings and sales, the company is not capping out. Unlike Apple (AAPL), which keeps recycling products, Nvidia (NVDA) keeps pushing the whole AI boundary forward.
Data center demand is still Nvidia's (NVDA) biggest driver. As robotics reaches full mainstream use - which may take a year or two - that revenue line should expand a lot, even if data centers cool off two years out. Nobody can forecast the exact growth, but either path points to a massive opportunity, so you get a two-for-one play. I have owned Nvidia (NVDA) since 2017, when it made gaming chips. It moved from gaming chips to crypto chips, then to GPUs, then to AI data centers, and next to robotics - executing all along the way.
The sensors: Analog Devices (ADI)
Analog Devices (ADI) fills the second layer, the sensing layer. When Optimus performed on stage at Tesla's (TSLA) investor day, tons of sensors made that possible, and that is Analog's business. It makes optic sensors and control systems that let a robot work reliably in the world with ultra precision. It is also inventing tactile sensing technology to turn commands into movement - replicating all the feedback your fingers get when you touch a mug.
Analog's (ADI) industrial segment sales approached $2 billion in the third quarter. EPS estimates sit at $16.66 for full year 2027, up sharply from this year, and jump to $19.66 for full year 2028. Robotics is a big driver, but the company is already executing on its existing businesses, so this is a large opportunity going forward.
The testing layer: Keysight (KEYS)
The least exciting layer, testing, makes sure all these systems work properly. Keysight (ticker KEYS) is a key name here. The critical question: you build a clever machine that does great things, but can it do them reliably without messing up? Keysight (KEYS) answers that. It has an AI software integrity builder introduced in 2026 that helps engineers validate AI models, test real-world behavior, and confirm it carries over.
Accuracy matters more as autonomous operations grow. People already use self-driving cars. Autonomous robots will take on serious work far beyond the cute restaurant robots and delivery bots - including medical tasks. That makes Keysight (KEYS) testing another strong opportunity.
