Manus and Genspark's global push shows Chinese AI agents competing outside China's firewall

Manus and Genspark, two of the most closely watched general-purpose AI agent startups to emerge from Chinese engineering teams, have each pursued aggressive international expansion strategies - domiciling holding structures outside mainland China, marketing primarily to Western and Southeast Asian users, and deliberately downplaying their Chinese engineering origins - reflecting a now-familiar playbook for Chinese consumer technology companies seeking global markets without the geopolitical friction that a visibly China-based AI product invites, particularly in the current environment of heightened scrutiny over Chinese technology companies' data-handling practices. The agent products themselves have generated genuine excitement for their ability to autonomously complete complex, multi-step tasks - researching a topic across dozens of sources and producing a structured report, planning and booking a multi-city trip, building and deploying a simple web application - with a level of end-to-end autonomy that has, in demo settings, impressed users accustomed to the more constrained, single-turn interactions of standard chatbots, even as independent testing has found significant reliability gaps between the most impressive demonstrated use cases and consistent real-world performance across the full range of tasks these agents claim to handle. The competitive response from American agent products - OpenAI's Operator and agent-mode ChatGPT features, Anthropic's computer-use capabilities in Claude, and Google's Project Mariner - has been to emphasise safety guardrails, enterprise trust credentials and integration with established productivity tools as differentiators against the more experimental, occasionally impressively autonomous but less predictable Chinese entrants, a positioning that reflects each side's assessment of where its actual competitive advantage lies. For Indian users and developers, both Manus and Genspark have found meaningful early adoption, particularly among small-business owners and freelancers drawn to the low-cost, high-autonomy value proposition for tasks like market research, content drafting and basic web development that would otherwise require hiring specialised help, even as data-privacy-conscious enterprise customers have generally remained more cautious about routing sensitive business data through agent products with opaque data-handling practices. What to watch: whether Manus or Genspark discloses funding or ownership details that clarify their relationship to Chinese parent entities more fully, how reliability benchmarks for autonomous multi-step agents mature to give users a clearer basis for comparison across products, and whether any Western government raises formal data-security concerns about these products' usage within regulated industries.

Original source: Rest of World