Nvidia, Microsoft and Google Announce New AI Chips and Agents for Laptops, With Caveats on Readiness and Cost
Nvidia introduced the RTX Spark chip on June 1 while Microsoft announced the Scout agent on Tuesday and Google previewed new Googlebooks features, all aimed at enabling AI agents to complete multi-step tasks locally.
EngadgetNvidia on June 1 revealed the RTX Spark chip for Windows laptops, designed to run AI agents without a cloud connection. The chip integrates Nvidia graphics, computing and networking technologies and carries more memory than standard laptop chips. Dell, HP and Lenovo plan to release computers using the new chip this fall.
Microsoft announced the Scout agent for Microsoft 365 on Tuesday. Scout incorporates OpenClaw technology and operates across cloud-stored content, local files and web data, including Outlook and Teams. The company said the agent is intended to monitor emails and work chats without repeated user prompts.
Google’s upcoming Googlebooks will suggest actions when a user points a mouse at items on screen, such as scheduling a meeting after hovering over a date in an email. Nvidia CEO Jensen Huang demonstrated a laptop running one of the new chips during a press conference this week, showing AI agents moving between 3D modeling applications to help design a house.
Bob O’Donnell, founder and chief analyst of Technalysis, said the goal is to let users tell a computer what they want and have the machine complete the task.
David Naranjo, associate director at Counterpoint Research, noted that familiarity with tools such as ChatGPT and Gemini has increased acceptance of AI agents. Jitesh Ubrani, research manager at International Data Corporation, said a range of issues must still be resolved before the technology reaches mass-market use.
OpenClaw, an AI assistant that gained popularity among developers this year, can execute programs and handle requests with limited prompting.
Developers have reported running the agent on dedicated machines while checking progress via WhatsApp or Telegram. Some employees have begun speaking commands to agents instead of typing, according to reports from Bloomberg and The Wall Street Journal.
ChatGPT launched in late 2022, after which large language models gained prominence and shifted development toward agents capable of multi-step work.
Earlier assistants such as Alexa and Siri handled single tasks like setting alarms but could not manage complex sequences or personal preferences. Naranjo said most consumers will not be able to control computers with a few prompts for some time, partly because the new laptops are expected to carry high prices.
Ubrani added that AI remains more useful for businesses than individual users and that local processing offers security and cost advantages for companies.
The announcements come after years of efforts by tech firms to move beyond simple voice commands toward systems that can plan and execute tasks autonomously.
Transparency
Reported by a single outlet. This score reflects source tier and factual specificity — corroboration is limited with one source.
Story details
Related Stories
Five Eyes agencies report Chinese military intelligence LinkedIn recruitment
Intelligence services from the U.S., Australia, Canada, New Zealand and Britain issued a joint alert stating that Chinese military intelligence is using fake profiles and job offers on professional networking sites to target government and military personnel with security clearan…
nbcnews.comTrump Signs Executive Order for Voluntary AI Model Testing
President Trump signed an executive order creating a voluntary 30-day government testing process for AI models. OpenAI CEO Sam Altman will meet with White House officials and congressional leaders on Wednesday to discuss the policy.
UK CMA Orders Google to Let Publishers Opt Out of AI Search Features
The Competition and Markets Authority gave Google nine months to implement new rules allowing publishers to block their content from AI-generated search results and AI model training.