New chips and agents promise local autonomy, but high costs and trust gaps could keep most consumers waiting for practical everyday use.

Global tech giants, including Nvidia, Microsoft, and Google, are actively investing in the development of autonomous AI management systems. Their chips, software, and platforms allow AI agents to act autonomously, solving complex tasks without constant human oversight.

Nvidia unveiled the laptop chip RTX Spark, designed to run AI agents without cloud connectivity, while Microsoft announced Scout for Microsoft 365.

Analysts warn that broad adoption faces obstacles such as high equipment costs and questions of trust, though the technology promises significant business opportunities.

“Ultimately, the goal is to figure out how to simply tell the computer what I want it to do, and then have it do it.”

– Bob O’Donnell

In May–June, companies continued to demonstrate how Windows updates can expand autonomous-agent capabilities. At the start of June, Nvidia unveiled RTX Spark – a Windows-based laptop chip that runs AI agents without cloud connectivity. The chip combines Nvidia’s graphics, computing, and networking technologies with more memory than typical laptops; device launches from Dell, HP, and Lenovo are expected this fall.

Plans are also emerging for Googlebooks – a future system that could propose actions when a user hovers the cursor over an object on the screen, for example to schedule a meeting after viewing a date in an email.

Prospects for Autonomous Computing and Real-World Trials

Tech companies have been trying for years to create digital assistants for everyday tasks: building shopping carts in online stores, tracking orders, and planning trips. Previously, assistants could perform single actions – calling a taxi or placing an order – but not multi-step tasks or understanding individual preferences. The shift began after large language models became central to developments following the launch of ChatGPT at the end of 2022.

OpenClaw – a somewhat under-the-radar AI assistant this year that has gained popularity among developers, can launch programs and handle queries without constant prompting. Previously, developers considered using it on home computers; as they move on to new tasks, they check the agent’s progress via messaging apps such as WhatsApp or Telegram.

Additionally, reports say that some tech workers have started giving commands directly to AI agents instead of typing them.

“Things are changing a lot right now because more people are getting used to services like ChatGPT, Gemini or Anthropic.”

– David Naranjo

Nvidia CEO Jensen Huang demonstrated at a press conference how a laptop powered by one of the new chips can help design a home, using AI agents to coordinate between 3D-modeling programs.

Microsoft is also working on a new agent for Microsoft 365 called Scout, using OpenClaw technology. Scout will be able to work with content in the cloud, on the user’s computer, and on the Internet, including Outlook and Teams apps. The company expects Scout to systematically monitor emails and business chats.

However, experts warn that pragmatism will often prevail: most people still won’t be able to control computers with only a few prompts over a long period, partly due to the high cost of new laptops. Despite rapid updates of agents from Anthropic and OpenAI, ordinary consumer adoption does not justify buying an expensive laptop to support the technology.

“For now this hasn’t become indispensable, and that’s where the challenge lies for Nvidia, Microsoft and others,” noted David Naranjo.

Artificial intelligence is usually more useful for business than for individual consumers, and performing tasks locally without sending data to the cloud can also be safer and reduce companies’ data-processing costs.

Before this becomes widespread, there is also a broader question of trust: are people ready to entrust their critical tasks to AI? If an agent buys tickets to a favorite performer’s concert, could it overshoot the budget?

“There is a whole spectrum of questions to answer before this becomes a mass market.”

– Jitesh Ubrani

“But are we on the right track? Yes.”

– Jitesh Ubrani

Despite the need for trust and the cost of equipment, the industry sees the potential of autonomous computing: the ability to perform important tasks without constant user involvement and reduced risk of data leakage that can occur when processing in the cloud. The business orientation of this technology promises advantages in customer service, manufacturing processes, and data protection.

Overall, analysts note that the path to mass adoption of autonomous computers is still not easy, but the race for more “self‑driving” systems continues, and success will depend on a combination of security, accessibility, and real usefulness for businesses and everyday users in daily life.