As AI proliferates, the edge case for the role of sensors is getting scrutinized by industry experts. Revenue potential already exists for sales of AI hardware—a collection of GPUs and CPUs and assorted accelerators used in hyperscalers.
But combining sensors with AI at the edge in industrial and work settings will also produce value beyond the hardware, and ultimately revenue for sensors companies themselves.
Some experts believe the next frontier for AI rests in sensory data collected from millions of devices—everything from what we already see in smartphones with accelerometers and other sensors—to sophisticated AI in motion for robots and cars.
That’s also the belief of Sanjay Kumar, who takes this prediction further. He’s a veteran of the silicon world of Intel, Broadcom and NXP who recently jumped from the US CHIPS program office to vice president of semiconductor and electronics at Kearney & Company, a government consultancy. Kumar will describe his views about AI sensors and more in a Sensors Converge 2025 keynote on June 26, “Fueling the AI Revolution: The Next Frontier of Data.”
“Why AI in sensors?” Kumar mused in an interview with Fierce Electronics. Data is being depleted for training AI models and that’s why synthetic data creation exists. “But we can train cars on the road and collect their data and use software to train the models with synthetic data. This opens the opportunity for AI sensor companies.”
For example, a sensor in a robotic arm operating in a work environment will collect all manner of data, such as how to position the arm to handle the weight of certain packages. Kumar envisions a memory component added atop the robot’s sensors, which will collect data that is later transferred to a company training foundational models. That data can be sold to companies developing physical AI models. “There could even be revenue generating opportunities, where companies providing sensors and companies using the sensors can cooperate,” he said. “The meta point is that sensors beyond hardware can do revenue generation themselves.”
For example, a company such as Schneider Electric could be installing sensors on the electric grid to watch for wildfires and then sell that data intelligence to an insurance company. “The hardware itself might not see a penny of revenue,” he said. Or, a Roomba vacuum cleaner in a home collecting data could help Amazon better understand its customers. “Potentially, the AI sensor hardware can see the benefit from data and the revenue will be case by case…I believe the hardware companies, like the Schneiders and others, are looking at it very carefully.”
One existing example of monetizing AI sensors beyond their hardware cost is how insurance companies are already urging drivers to put safe driving sensors inside their cars to qualify for cheaper insurance. “That use case is already there,” he said, but how the savings is being distributed in the value chain will vary. Also, dozens of different sensors are already being used to add value to autonomous vehicles and robotaxis. “The guy making sensors is able to generate more business,” he noted.
The arrival of robotaxis in China and, lesser so, in the US has shown the value of the sensor relationship to robotic AI. “I’m surprised how quickly the robotaxi is happening,” with Waymo on the roads in a few cities, he said, after having worked on autonomous vehicles years ago. “Elon Musk overpromised, for sure,” although reports indicate Tesla will unveil its robotaxi in operation in Austin, Texas, by June.
Kumar predicted the future for automobiles “will not be fully autonomous for sure, but good enough for 80%” of driving scenarios. “That last 20% is the hardest and would require a computer that’s exponentially more expensive.”
Asked if the current state of tariffs on electronics is a matter of great concern, he answered, “Absolutely. Tariffs are going to ripple through the whole economy. Electronics infrastructure is the foundation of everything in the modern world. No supply chain is self-sufficient. We are dependent on other parts of the world.”
Some companies will eat all the costs associated with tariffs, but others will pass them on to customers. “Electronics will become expensive. Even a toaster, which is not complex technology with a simple heat sensor, and other things are going to get expensive. That part of the economy is not very high margin. Costs will get passed on to consumers.”
Sanjay Kumar will deliver the keynote at 9 a.m. PT on Thursday, June 26, at Sensors Converge 2025 in Santa Clara, CA. Go online to register and obtain a free Expo Hall pass with the code HAMBLEN.