China’s Manus AI Could Redefine Autonomous Agents, Experts Say
China’s AI industry has unveiled what many experts are calling a potential game-changer in artificial intelligence. Manus AI, developed by startup Monica.im, claims to be the world’s first fully autonomous AI agent, capable of independently executing complex tasks from start to finish with minimal human guidance.
The development comes just months after Chinese AI company DeepSeek sent shockwaves through the tech industry with its powerful language model that matched capabilities of systems costing billions to develop.
“This isn’t just another chatbot workflow, it’s a truly autonomous agent that bridges the gap between conception and execution,” said Wes Roth, a prominent AI industry analyst who was among the first to report on Manus. “While other AI stops at generating ideas, Manus delivers results.”
Early demonstrations show Manus performing diverse tasks that would normally require significant human effort — screening job applicants by analyzing resumes, researching real estate markets, conducting financial analysis, and even creating functional websites without step-by-step human guidance.
The system achieves this through a sophisticated technical architecture that coordinates specialized AI components working in concert: planning agents that interpret user goals, execution agents that carry out tasks, and verification agents that validate results.
“The launch of DeepSeek was a massive moment for the American AI industry,” Roth noted. “It was described as a Sputnik moment for US AI development. Now, not that long after DeepSeek, we have Manus.”
In my recent San Francisco Chronicle op-ed about DeepSeek’s impact, I wrote that “DeepSeek R1 delivers Chat GPT-4-level performance at a fraction of the cost, obliterating a central assumption of modern AI: that only billion-dollar tech giants could build frontier models.”
According to published benchmarks, Manus has achieved state-of-the-art performance on the GAIA benchmark, outperforming OpenAI‘s models across all difficulty levels, including a significant lead in the most complex tasks.
Perhaps most notably, Monica.im has committed to open-sourcing components of Manus later this year, potentially accelerating innovation in AI agent technology globally. This approach contrasts with the increasingly closed, subscription-based models of Western AI companies like OpenAI, which is reportedly discussing AI models priced at $2,000 per month, with potential plans for even more powerful systems at around $20,000 monthly.
The timing is significant as OpenAI recently unveiled their own autonomous agents, Operator and Deep Research. While these represent significant advancements, Manus appears to offer potentially greater autonomy with the added benefit of planned open-source components.
Currently in an invitation-only preview phase, Manus has reportedly generated overwhelming demand that has strained server capacity. “The people that do have their hands on it, that have tested it out, seems like they’re kind of going crazy for it,” Roth said.
As I concluded in my Chronicle analysis, “We’ve crossed a threshold. The old playbooks — whether for business strategy, national policy or career planning — are obsolete overnight.”
The tech industry now watches closely to see if this “DeepSeek 2.0 moment” will fundamentally redefine our expectations of what AI systems can accomplish without human supervision.
