What's the bet behind the multi-billion-dollar acquisition? N key interpretations of Manus siding with Meta
On December 30th, Beijing time, Meta announced the completion of its acquisition of the general autonomous AI agent company Manus (Butterfly Effect). The transaction amount is approximately billions of dollars.
The most frequently used word in the comments from the domestic tech circle about this transaction is "shocked".
After Tencent Technology communicated with multiple investors and creators in the Agent field, the shock mainly focused on the following points:
Manus was recently reported to be seeking the next round of financing at a valuation of $2 billion. Unexpectedly, it was acquired by Meta so soon.
The description of "billions of dollars" leaves a lot of room for imagination. If it is over $4 billion, Meta has truly shown great sincerity (anxiety).
Why was a startup founded in China ultimately acquired by Meta?
But there are also several points of consensus:
The selling price is very good. The founding team of Manus made several key and excellent choices. We are happy for Manus.
After all, general agents are on the "main track" of large models, and being acquired by a big company is their destined fate.
Manus has strong product capabilities and good reputation and popularity in Silicon Valley, which is also a solid foundation for its valuation.
This is the best outcome and the best timing.
At the end of 2025, the combination of MM (Manus - Meta) may, just like the company name of Manus - "Butterfly Effect", cause ripples in the tech circle.
01 How much is Manus worth?
According to the official announcements and public information of Meta and Manus, the transaction amount reached billions of dollars, making it the third - largest acquisition in Meta's history, second only to the acquisition of WhatsApp and the previous strategic investment in Scale AI.
According to the information disclosed by both parties, after the completion of the transaction, Manus will continue to operate relatively independently. Xiao Hong, the founder and CEO of the company, will serve as the vice - president of Meta, responsible for leading Meta's global business in the direction of general - purpose AI agents.
Multiple people close to the transaction revealed that the acquisition negotiation between Meta and Manus progressed extremely fast, lasting only more than ten days from the initial contact to the final signing.
How much is Manus really worth? Based on comprehensive analysis of multiple sources of information, the acquisition price may be in the following range:
The most likely range: $2 - 3 billion
Valuation premium multiple: A 25% - 50% premium is given on the basis of the financing valuation ($2 billion).
Revenue multiple: $2 - 3 billion corresponds to 16 - 24 times the ARR, which is in line with the valuation level of high - growth SaaS companies.
Strategic value: Considering Manus's technological leadership, paying user base, and rapid growth potential.
Transaction speed: Completing the negotiation in more than ten days indicates that both parties quickly reached a consensus on the price range.
From the lower limit of the valuation, the price of this transaction is about $2 billion because before Meta's intervention, Manus was just reported to be preparing for a new round of financing at this valuation, and foreign media generally believe that based on the previous valuation of $500 million, $2 billion is a logical starting point for the premium.
From the upper limit, the acquisition amount should not exceed $5 billion, mainly because Manus's annual recurring revenue (ARR) of $125 million is limited in scale and difficult to support a higher valuation multiple. As the third - largest acquisition in Meta's history, its price should be significantly lower than the previous two - Scale AI (about $14.3 - 15 billion) and WhatsApp (about $19 billion).
An investor focusing on early - stage venture capital told Tencent Technology that based on the current ARR of $125 million, compared with current Silicon Valley startups, a valuation of $2 - 3 billion is reasonable. If it is $4 - 5 billion, the valuation is on the high side.
02 Meta: Is Manus indispensable?
Meta acquired Manus at a not - low price and without hesitation like lightning. According to public information, "the negotiation was completed in more than ten days."
So, what did Meta value in Manus?
Meta officially positioned Manus as a key "execution - layer" component in its AI strategy. The goal is not to launch a single product but to gradually integrate this intelligent agent system with end - to - end execution capabilities into its huge social and enterprise service platforms.
From a business perspective, Manus achieved an annualized revenue (ARR) of over $125 million within only eight months of its launch, demonstrating an extremely rare commercialization speed.
Its technical solution has also been verified under high loads in the real world: the system has cumulatively processed more than 147 trillion tokens and supported the operation of more than 80 million virtual computers. This engineering resilience means that Manus is no longer a demonstration model in the laboratory but has the ability to provide continuous and reliable delivery services for global users.
More deeply, it also reflects Meta's anxiety about transformation.
A startup founder in the Agent field in Silicon Valley told Tencent Technology: "To understand this, we need to put it in the big picture. We can compare the current AI era to the early days of the mobile Internet. Meta has once again resorted to the same strategy as when it acquired Instagram and WhatsApp: in the critical window period, 'buying' is definitely more efficient than 'building and operating on its own'."
Meta's current situation is quite urgent. Its original main story, Llama, is showing signs of fatigue and can't compete with its opponents. If it can't come up with something like Manus that may break the deadlock, Wall Street will quickly classify it as an "old - generation" company. So, the cost - effectiveness of this transaction is not important at this moment. Instead of letting the money sit on the books as a floating figure in the secondary market, it's better to quickly turn it into an asset like Manus. This is the key for Meta to survive and turn the tables in the AI race."
So, does Meta have any other better choices besides Manus? "Perplexity was once an option, but at that time, Meta's entire strategic focus was on Llama. Now, Manus, which has both a mature product and a user base, may be the best choice," an investor in the tech field told Tencent Technology.
03 Manus, the best outcome?
Meta needs Manus. So, is being acquired by Meta the best outcome for Manus? On this issue, the views of multiple investors and AI - field startup founders are surprisingly consistent.
"This is the best outcome and the best timing."
"Xiao Hong, the co - founder of Manus, has a strong product intuition. And in the development of Manus, he made several very decisive and accurate choices," an investor close to Manus told Tencent Technology.
The several key choices made by Manus include:
1. Although the company (Butterfly Effect) was founded as early as 2022, what really changed its fate was in October 2024 when the team decided to develop the first general - purpose Agent product, Manus, and created a growth miracle of exceeding $100 million in ARR within eight months.
2. Among them, Manus made an extremely difficult decision: even though the founding team and most of the R & D forces are located in China (Beijing, Wuhan), it chose an overseas first - launch when releasing the product and used top - level models such as Anthropic, which are inaccessible in China, as the underlying support. Although facing huge public pressure, it exchanged for the explosion of the global market.
3. After receiving a $75 million financing from Benchmark, Manus quickly moved its headquarters to Singapore, clearing the biggest legal and capital obstacles for its subsequent acquisition by a US giant like Meta.
4. This time, without hesitation, it accepted the olive branch extended by Meta instead of insisting on "going public independently", which was another key choice.
Although Manus's product is powerful, general agents are still on the main track of large models.
A Silicon Valley startup founder told Tencent Technology: "I conducted an early evaluation of Manus. Its technical barriers are not strong, but the product iteration is very fast. Xiao Hong has a very good product intuition. Being acquired by a big company is the best outcome and a win - win situation, but it also exposes the bottleneck of AI startups: Profit Margin."
For a general - purpose Agent product like Manus, every task execution requires the use of expensive large models and cloud virtual machines. Although its ARR (annual recurring revenue) has exceeded $125 million, if the bills paid to cloud providers (such as AWS, NVIDIA) and underlying model companies are deducted, the remaining "net profit" may be much thinner than that of traditional software companies.
Under this business model, the survival space and price - raising ability of startups will be extremely compressed. When Manus only had 10,000 users, the losses were controllable, but as the user scale continues to grow, "Profit Margin" becomes an even more difficult bottleneck to cross.
Manus's subscription price is basically the same as that of global SOTA large models. If it can't clearly prove its superiority over "general large models", large model companies with more funds and user bases obviously have stronger competitiveness.
"Now is the best timing. If it doesn't sell now, it may not be able to sell later or may not get such a good price." In terms of timing and price, both Manus and its investors have received generous returns.
For Manus: In less than three years since its establishment, its valuation has jumped from $14 million to $2 - 3 billion. For investors: They have obtained a considerable return of 5 - 12 times. For example, ZhenFund, as an investor in the seed and angel rounds, participated in all four rounds of financing and is expected to obtain an overall return of more than 10 times, verifying the accuracy of its early judgment.
However, Manus's acquisition has once again sparked controversy: Why did Manus sell to Meta instead of a Chinese company?
One view is that domestic investors hesitated and didn't understand the underlying logic of Manus. Another view is that the valuation was too high, the cost - effectiveness was not good, or they simply didn't have the money to invest.
"At that time, except for Sequoia, it can be said that all domestic investors rejected Xiao Hong," a person close to Manus told Tencent Technology. "After Manus took shape, it was Benchmark. The valuation offered by domestic capital was not high enough."
Benchmark Capital is one of the most legendary venture capital institutions in Silicon Valley. In April 2025, Benchmark led Manus's Series B financing of $75 million. If Meta acquires it for $2.5 billion, Benchmark will obtain a return of about 5 times in just eight months.
Regarding today's outcome, the aforementioned person close to the matter said: "It's good. Both investors and entrepreneurs have received returns."
Regarding the future, there are different judgments in the industry on whether Manus's founding team will continue to stay at Meta and whether Xiao Hong will really remain as the vice - president of Meta.
One view is that Meta also bought the people of the Manus team, and they won't leave in the short term. Another view is that for the Manus team, this round of entrepreneurship has ended perfectly. As the VP of Meta, Xiao Hong's position is just a symbolic acquisition condition.
Some investors gave a more macroscopic evaluation of this event: "This acquisition is regarded as a milestone in the AI Agent industry and a symbol of the global competitive spirit of a new generation of Chinese entrepreneurs."
This article is from the WeChat official account "Tencent Technology", author: Guo Xiaojing, editor: Xu Qingyang. Republished by 36Kr with permission.