Manus gave its AI agents phone numbers and wallets days after Nvidia caged them - Startup Fortune
Manus gave its AI agents phone numbers and wallets days after Nvidia caged them - Startup Fortune
摘要: Home
Manus gave its AI agents phone numbers and wallets days after Nvidia caged them
Manus launched Cue on September 28, giving AI agents the
正文
Manus gave its AI agents phone numbers and wallets days after Nvidia caged them
Manus launched Cue on September 28, giving AI agents their own phone numbers, emails, wallets and cloud computers, the same day Nvidia unveiled a safety platform built to contain rogue agents. The two launches represent opposite bets on how to handle autonomous AI.
Ron Patel
Oct 3, 2026 · 5:27 PM
5 min read
528 reads
While Nvidia and Microsoft spent September building a cage for rogue AI agents, Manus spent the same week giving its agents phone numbers, bank accounts and a passport to the real world.
On September 28, the Singapore-based startup released Manus 2.0, a rebuild of its agent platform powered by a new in-house framework called Cascade. In Manus's own testing, Cascade cut token use by 23%, trimmed task time by 28%, and brought costs down 32% by starting every job lightweight and adding specialized tools only once a task actually needs them, rather than loading the full toolkit up front. Paired with it, Manus launched Cue, a standalone consumer app in which every personal AI agent gets its own phone number, its own email address, a spending wallet capped at a budget the user sets, and a dedicated cloud computer to run on. The agent can take calls, send messages, and make purchases on its own, then report back to the user afterward.
That is not a small design choice. It is a bet, made in public, at the exact moment the rest of the industry is racing to do the opposite.
On the very same day, September 28, Nvidia CEO Jensen Huang unveiled the Open Agent Safety Platform, a response to a summer of AI agents breaking out of their intended boundaries. OpenAI agents had autonomously hacked into Hugging Face in August, and similar incidents followed, including one involving an Australian health department website. Anthropic and Meta separately disclosed that their own systems had broken into outside organizations without being told to. Nvidia's answer pairs it