Summary

Cerebras Systems CEO and co-founder Andrew Feldman believes personal AI assistants could soon become a regular part of everyday life. Speaking at TechCrunch Disrupt 2026, Feldman predicted that people may eventually use AI appliances that combine on-device processing with cloud-based services. These assistants could learn users’ preferences, manage daily tasks, access personal information, and even make purchases with permission. His comments highlight the growing interest in AI agents that can do more than answer questions.

What Happened?

Andrew Feldman, CEO and co-founder of AI chipmaker Cerebras Systems, shared his vision for the future of consumer AI at TechCrunch Disrupt 2026.

During the discussion, Feldman was asked how long it might take before AI becomes a common household appliance. He suggested that this future may be closer than many people expect.

Rather than relying entirely on cloud-based chatbots, Feldman believes people could have personal AI assistants that operate partly on their own devices and partly through cloud services.

The idea is to create assistants that understand individual users over time and help manage practical responsibilities, from organising schedules to finding information and completing everyday tasks.

How Would a Personal AI Assistant at Home Work?

Feldman expects future personal AI assistants to use a combination of local computing and cloud infrastructure.

Some tasks would run directly on a device inside the home, while more demanding requests could be processed through cloud-based AI services. This hybrid approach is similar to how many modern smartphones combine on-device features with online services.

A personal AI assistant could potentially help users with tasks such as:

  • Managing schedules: Reviewing calendars, preparing for meetings, and reminding users about important events.
  • Finding relevant information: Gathering details about people, appointments, travel plans, or upcoming activities.
  • Handling routine tasks: Helping organise household responsibilities and everyday digital activities.
  • Making purchases: Ordering products or services when the user has granted the necessary permission.
  • Learning personal preferences: Adapting its responses and recommendations based on the user’s needs and habits.

These capabilities would depend on the software, available integrations, and permissions provided by the user. A personal AI assistant would not automatically have access to financial accounts or the authority to make purchases.

Why Running AI Locally Matters

One important part of Feldman’s vision is that personal AI assistants could perform some work directly on a device rather than sending every request to a remote server.

Local processing may offer several advantages, including faster responses for supported tasks, reduced dependence on an internet connection, and greater control over certain types of personal data.

However, local AI does not automatically guarantee privacy or security. Devices would still need appropriate safeguards, and cloud-based requests could involve sending information outside the home.

For assistants that manage calendars, personal contacts, financial details, or shopping accounts, protecting sensitive information will be especially important.

Feldman believes users could eventually have a security boundary around their personal assistants. Making that vision practical will require strong access controls, clear permissions, and ways for users to review or stop actions performed on their behalf.

AI Agents Are Moving Beyond Simple Chatbots

Feldman’s comments come as technology companies increasingly explore AI agents that can carry out multi-step tasks instead of simply responding to questions.

Traditional chatbots primarily generate answers based on a user’s prompt. AI agents are designed to take additional steps, such as gathering information, interacting with software, and completing a task through connected services.

Feldman pointed to developments in consumer AI agents as examples of how the technology is moving toward more practical everyday uses.

If these systems become dependable, users may be able to delegate more routine digital work to AI rather than managing each step themselves.

However, agents still face challenges involving accuracy, reliability, security, and the ability to handle unexpected situations. Tasks involving payments or sensitive personal information will require particularly careful oversight.

Cerebras CEO Already Uses AI Agents for Daily Work

Feldman also explained that he personally uses a collection of AI agents to help manage his working day.

According to his comments, these tools can review his calendar and gather information about people he is scheduled to meet. Such assistance can reduce the time spent preparing for calls and searching across different sources.

For Feldman, the most exciting aspect of AI is not simply automating activities people already perform. He is also interested in applications that make entirely new capabilities possible.

He said he prefers working on the underlying infrastructure that enables other companies and developers to create new products, rather than trying to predict every successful consumer application himself.

That perspective reflects Cerebras Systems’ position in the AI industry. The company develops computing hardware designed to support demanding AI workloads, while software developers and product companies build the applications people use.

What Could This Mean for the AI Industry?

A shift toward personal AI assistants could create new opportunities across several parts of the technology industry.

Hardware manufacturers may develop devices designed specifically for running AI locally. Software companies could build assistants that connect calendars, email, shopping services, and smart-home systems. Cloud providers would continue to play a role in supporting more demanding AI operations.

For consumers, the potential benefit is convenience: an assistant that can understand context and carry out useful tasks with less manual input.

At the same time, widespread adoption will depend on whether these assistants can earn users’ trust. People will want to know what information an assistant collects, where that information is processed, and how much control they retain over its actions.

The ability to prevent unauthorised purchases, protect personal data, and correct mistakes will be just as important as the intelligence of the underlying AI model.

When Will Personal AI Assistants Become Common?

Feldman believes the idea is not far away, but his comments should be understood as a prediction rather than a confirmed product launch or industry-wide timeline.

Some elements of his vision already exist in current AI services, including scheduling assistance, conversational interfaces, local AI processing, and tools that can interact with connected applications.

Combining these capabilities into a dependable home appliance that securely manages many aspects of a person’s life remains a more ambitious goal.

Progress will depend on improvements in AI models, affordable hardware, software integrations, privacy protections, and consumer confidence. It remains unclear when fully capable personal AI appliances will become commonplace.

Conclusion

Cerebras CEO Andrew Feldman believes personal AI assistants could become a familiar part of the home, helping people manage information, organise their schedules, and complete everyday tasks.

His vision combines local processing with cloud-based AI, potentially giving users assistants that understand their preferences while maintaining a secure boundary around personal information.

Although the timeline remains uncertain, the growing development of AI agents suggests that the technology industry is moving beyond simple chatbots toward systems that can take action. The key challenge will be making these assistants useful, reliable, secure, and worthy of users’ trust.