Personal AI assistants could soon move into the home, according to Cerebras Systems CEO and co-founder Andrew Feldman. Speaking at TechCrunch Disrupt 2026 on Tuesday, he described a future in which assistants combine local computing with cloud access, learn about the people who use them, and handle tasks such as buying things.
Feldman sees home AI approaching
Asked what would need to happen for AI to become a home appliance, Feldman replied, “Actually, I don’t think that’s very far away.” His answer was a prediction, not a report that such household assistants are already widely available.
He pointed to advances in consumer AI agents as signs that the technology is finding its footing with everyday users. The examples he named were Meta’s Muse and Instinct, which he described as a startup now valued at $10 billion.
The broader idea is that an AI assistant could become a regular part of life at home. Feldman did not set out a launch date or describe a particular appliance. Instead, he outlined how he expects personal assistants to work and what they might be able to do.
A split between the device and the cloud
Feldman expects assistants to run locally while also reaching back to the cloud. He compared this arrangement to a cell phone: some work would happen on the device, while other work would go back to the cloud.
That split is central to his picture of a home assistant. The assistant would not depend on only one place to do its work. Feldman’s description leaves the specific division of tasks open, but it suggests a system designed to combine computing in the home with cloud-based resources.
He also expects these assistants to get to know their users well. Alongside that personal knowledge, he said, they would have “a security perimeter around them.” The source does not explain how that perimeter would be built or what protections it would include, so those details remain part of the vision rather than a specified design.
From information gathering to purchases
Feldman’s forecast includes more than answering questions or finding information. He said personal assistants would “certainly have access to credit cards and be able to buy things for you.” That would put purchasing among the tasks he expects an assistant to handle.
His comments also make clear that this capability depends on assistants knowing their users and having a security boundary around them. Those ideas appear together in his description of home AI: personal context, access to useful services, and the ability to act on a user’s behalf.
Feldman believes the result could make “a meaningful impact in making our lives better.” That is his expectation for the technology’s effect. He did not identify specific improvements or say when people would experience them.
How Feldman uses AI today
Feldman said he already uses “a collection of agents” in his own work. He uses them to sweep his calendar, pull up information about people he is going to have a call with, and handle more tasks.
Those examples show current uses he described, while the home assistant is a prediction about what may come next. He said AI is already doing many things people used to do manually, but his strongest interest is in possibilities that were not possible before.
He left it to other founders to discover what those new possibilities might be. Feldman said he likes working on infrastructure that lets other people’s ideas take flight, rather than needing to develop the consumer-facing idea himself.
That distinction connects his personal use of agents with his role at a chipmaker: he can see potential in the applications while focusing on the underlying platforms. Whether home AI reaches the form he described remains uncertain, but his forecast puts local computing, cloud access, personal knowledge, security, and purchasing in the same picture of a future assistant.