AI agent developer Manus completes a new financing round at a $4 billion valuation just seventeen days after its relaunch.

Manus, barely over half a month into its restored independence, has seen its valuation suddenly double. The company is currently advancing a new financing round of approximately $500 million, targeting a valuation of around $4 billion. This marks its first fundraising effort since completing its separation from Meta and resuming independent operations. According to Bloomberg citing informed sources, the round is nearing completion, though negotiations remain ongoing and final amounts and terms are subject to change. The identity of the latest investor has not been disclosed, and existing shareholders—including Tencent, HSG, and ZhenFund—have not publicly confirmed whether they will increase their stakes.

On September 1, Manus officially announced its return to independent operations. Less than a month earlier, former shareholders such as Tencent, ZhenFund, and HSG repurchased the company from Meta for roughly $2 billion—the same price Meta had paid during its initial acquisition. If this latest round closes at the $4 billion valuation, Manus’s post-independence valuation will have doubled in just 17 days. Over the past year and a half, Manus’s valuation has climbed steadily from $500 million to $2 billion, and now targets $4 billion, navigating through an acquisition, regulatory intervention, business divestiture, and a shareholder buyback. Undoubtedly, it stands out as one of the most dramatic agent companies globally.

Bloomberg initially reported that Manus is seeking approximately $500 million in funding at a $4 billion valuation. Notably, it took only 18 months from the product’s initial launch to reach this $4 billion mark, a trajectory punctuated by several high-profile twists. In March 2025, Manus launched its general-purpose agent, generating immediate industry attention. By April, Benchmark led a $75 million investment round, placing the company’s post-money valuation at roughly $500 million. At the end of December, Meta announced a $2 billion acquisition of Manus, which had just declared an annual recurring revenue (ARR) exceeding $100 million and boasted millions of users, with an annualized run rate of $125 million. However, the cross-border deal was unwound in early 2026, prompting Meta to begin separating Manus’s business and data assets.

In May, Bloomberg reported that founders Xiao Hong, Ji Yichao, and Zhang Tao had explored raising approximately $1 billion to repurchase Manus from Meta, requiring a valuation of at least $2 billion. They also considered operating through a joint venture structure before eventually pursuing a Hong Kong IPO. That plan did not materialize. Instead, in August, original investors including Tencent, ZhenFund, and Sequoia China acquired Manus’s shares from Meta for roughly $2 billion. According to Caixin, Tencent assumed Benchmark’s previous stake, becoming Manus’s largest shareholder. (Notably, founder Xiao Hong has longstanding ties to Tencent; his early WeChat management tool, “Yiban Assistant,” was previously backed by Tencent.) On September 1, Manus formally announced its independence, with the founding team reinstated and the company repositioned as an “independent Agent laboratory.” Just 17 days later, the $4 billion valuation proposal emerged.

While Manus is not the only company to experience a failed acquisition followed by a shareholder buyback, most such cases see valuations stagnate or decline post-buyback, often as a loss-mitigation measure rather than a value reassessment. Only rare instances of explosive organic growth achieve higher valuations afterward, such as Figma, which listed at a premium to Adobe’s abandoned acquisition offer. Manus clearly falls into the latter category. Despite the turmoil, Manus’s business metrics tell a compelling story. Prior to Meta’s December acquisition, Manus reported ARR surpassing $100 million. By late June, according to Caixin, ARR had reached approximately $400 million. The Information cited sources indicating an even higher figure, with Manus’s annualized run rate hitting $400 million to $500 million in June. Using the publicly reported figures, Meta’s $2 billion purchase in December corresponded to roughly a 20x ARR multiple. Based on the June $400 million ARR, the current $4 billion target implies a multiple of approximately 10x ARR. Thus, while the absolute valuation doubled, the valuation per dollar of ARR effectively halved. When the company was bought back in August, revenues were already significantly higher than at the time of Meta’s acquisition, meaning original shareholders effectively reacquired a larger, more profitable entity at the old valuation. The newly disclosed $4 billion figure appears to be the market pricing in that operational growth.

Beyond revenue expansion, Manus has maintained rapid product iteration throughout the separation period. Its development roadmap has consistently focused on connecting additional tools, preserving broader project context, enabling external tool invocation, and extending task execution from minutes to hours, days, or continuous operation. Public update logs from spring and summer show steady progress along these lines. Even amid the Meta divestiture, Manus continued advancing partnerships with Canva, Zoom, Similarweb, Notion, and Shopify. Upon announcing its independence, Manus outlined its next phase: deeper integration into daily workflows, more direct interaction with the external environment, and increasingly proactive action on behalf of users. These factors clarify the rationale behind the $4 billion ask. Subscription revenue, workflow entry points, user accumulation, and the potential to secure a lasting product layer atop foundational models are all factored into the valuation.

Crucially, beyond these fundamentals, Manus holds a strategic position as a leading player in this year’s hottest agent sector. Before this year, Manus was among the few globally available general-purpose agents with notably strong user experience. Since January, competitors like OpenClaw have surged in popularity, while Claude Code and Codex continuously expand their agent capabilities. Domestic developers are similarly accelerating progress in long-horizon tasks, tool use, and autonomous execution. Furthermore, foundation model providers themselves are increasingly building native agent functionalities. Each major model update risks absorbing capabilities from upper-layer agent products. Manus does not train frontier foundational models, placing it in a highly competitive environment with established giants on both sides. Nevertheless, its current revenue trajectory demonstrates that, as of today, startups can still build valuable, rapidly growing agent products atop other models without developing base models themselves.

Whether the $4 billion valuation will find buyers remains to be seen. Three critical questions loom over Manus. First, can the approximately $400 million ARR continue its rapid growth? While the jump from $100 million to $400 million took only six months, sustaining such momentum becomes increasingly difficult, especially as top-tier model vendors bundle agent features directly into their offerings. Manus must continually prove its subscription value remains distinct. Second, who will fund the $500 million round? Will existing shareholders reinvest? Will new large-scale U.S. dollar funds, industrial capital, or sovereign wealth vehicles enter? The investor list remains undisclosed, and its revelation will clarify the capital logic behind the valuation. Third, what will Manus evolve into post-independence, and how will it steer product strategy? Will it become a super-agent spanning multiple scenarios, or a lightweight agent workspace for teams and small-to-medium enterprises? Answers remain pending. Regardless, having navigated significant turbulence, Manus retains a strong hand. In an era poised for broader agent adoption, it has already captured first-mover mindshare.