Manus

1. 2025-03

   Launch; invite-code scarcity

   Codes resold second-hand; two sources give different prices

   [7]

2. 2025-03-28

   First paid plans (Sacra)

   "The company introduced its first paid subscription plans in March 2025"

   [8]

3. 2025-07

   Layoffs in the China office, team to Singapore; Wide Research launch

   [7]

4. 2025-12-17

   $100M ARR, $125M run-rate

   "Manus has crossed $100M in ARR, eight months after launch"

   [1]

5. 2025-12-29

   "Manus Joins Meta"

   "without changing how Manus works or how decisions are made"

   [2]

6. 2026-04-27

   China blocks the acquisition

   "required all the parties to withdraw from the deal"

   [3]

Acquisition channels

At launch, invite codes became the coveted object on X and Chinese social media; according to Peak Ji the scarcity was not marketing but a "capacity constraint" [6]. The waitlist passed 2 million before monetization (Sacra, secondary data) [8]. The second channel was the "first general agent" positioning; the $100M post refers to the category "as we defined this category together" [1]. The third was release cadence: Wide Research, 1.5 and 2.0 each worked as a relaunch [5]. The fourth was the Microsoft Agent 365 partnership (November 2025) [5].

Loop

Scarcity fed social curiosity, curiosity fed the waitlist, and the waitlist fed the paid plan of March 2025 [8]. As agent outputs (sites, reports) were shared, new tasks and credit consumption followed. In December 2025 the company claimed month-over-month growth above 20% [1]. A second loop runs through enterprise partnerships and team plans [5].

Activation

The invited user's first task. (opinion) The observable virtual computer, watching the screen while the agent works, is the first value moment.

Retention evidence

No official retention figure. 36kr wrote that user feedback declined significantly after May 2025 [7]; yet in December 2025 the company claimed over 20% monthly growth [1]. (opinion) The launch wave and the product-market-fit wave are different things; between them sit the relocation and the repricing. No retention evidence.

Role of pricing

A task is a credit. The company reported "ARR" and "run-rate" separately: $100M in subscriptions, $125M including usage [1]. (opinion) That split is a rare piece of transparency among AI companies.

Distribution moves

Release cadence and platform partnership: Wide Research on 31 July 2025, Manus 1.5 on 16 October 2025, Microsoft Agent 365 on 19 November 2025, Manus 2.0 on 28 September 2026 [5]. The Meta announcement said subscriptions and the Singapore operation would continue [2].

Team

105 employees in December 2025 across Singapore, Tokyo and San Francisco (company blog) [1]. Layoffs in the China office in July 2025 [7]. (opinion) $100M ARR over 105 people is roughly $0.95M of annual revenue per employee, a ratio computed from two separate sources.

The five questions of the Growth Claim Check, applied to this company's public claims.

Test your own claim with the same five questions: Growth Claim Check →

• Four months after the Meta announcement, China's NDRC prohibited the deal and told the parties to withdraw [3]. Meta said the deal "complied fully with applicable law" [4]. The Manus 2.0 announcement suggests the company continues independently [5]; the ownership structure could not be verified.
• User feedback fell after launch; layoffs in the China office and the move to Singapore followed in July 2025 [7].
• Invite-code black market: 36kr reports prices above ten thousand dollars, another source above one thousand; both secondary [7].
• No foundation model of its own; which models are used, and the margin, could not be opened from a primary source on research night.

This section is opinion; a reading, not a source.

1. 01

   (opinion) Scarcity only works when the capacity constraint is real. Artificial scarcity is seen through once; Manus's scarcity was real and feedback still fell two months after launch.

2. 02

   (opinion) State ARR and run-rate separately. Splitting subscription from usage revenue built trust; a founder in Turkey who says "ARR" should make the same split.

3. 03

   (opinion) A launch wave is not product-market fit. The second growth came from relocation, repricing and release cadence.

4. 04

   (opinion) Geography is a risk line item. A Chinese-origin company's sale to the US was stopped by the state; for a company founded from Turkey, data residency and ownership structure are early decisions.

5. 05

   (opinion) A platform partnership accelerates; it is not a moat. Test question: if Microsoft promotes its own agent tomorrow, why does your customer stay?

A question for you: remove this company's logo and drop the loop onto your own product. Which node stays empty?

Numbers are the references in the text. An 'opened' source was read by the researcher; 'secondary' ones are verified against their primary source before publication.

1. [1]
   Manus Update: $100M ARR, $125M revenue run-rate

   Manus blog · 2025-12-17 · opened

   “Our company's total revenue run rate is now over $125M, including usage-based and other revenue.”

2. [2]
   Manus Joins Meta for Next Era of Innovation

   Manus blog · 2025-12-29 · opened

   “Joining Meta allows us to build on a stronger, more sustainable foundation without changing how Manus works or how decisions are made.”

3. [3]
   China blocks Meta's acquisition of Manus

   RTHK · 2026-04-27 · opened

   “required all the parties to withdraw from the deal”

4. [4]
   China blocks Meta from acquiring AI startup Manus

   AP (via NY1) · 2026-04-27 · opened

   “complied fully with applicable law”

5. [5]
   Blog index (release dates)

   Manus · accessed 2026-10-05 · opened

   “Introducing Manus 2.0”

6. [6]
   'Mens et Manus', the story of Manus

   AI Proem · 2025 · opened

   “Everything added dilutes everything else”

7. [7]
   Manus 'Comes Ashore': What You Need to Know

   36Kr (EN) · 2025 · opened

   “After May, as more users got the test qualification, the word - of - mouth took a nosedive.”

8. [8]
   Manus profile (estimates)

   Sacra · 2026 · opened

   “The company introduced its first paid subscription plans in March 2025”

9. [9]
   China Blocks Meta's $2 Billion Acquisition of AI Firm Manus

   Bloomberg · 2026-04-27 · secondary

   Note: headline only; text not opened

• Ownership after the blocked deal: does Meta hold a stake, was money returned?
• A primary Reuters or Bloomberg text for the Benchmark round.
• The official pricing page; Sacra tiers need verification.
• Second-hand invite-code price: one thousand or ten thousand dollars; both secondary.
• Which foundation models are used; margin.
• The exact launch date and the official launch post.