Meta just acquired a Chinese-founded AI startup for $2 billion. Here's why it's important

AI For Business


A darling of the artificial intelligence startup scene has been acquired by Meta Inc. The move capped a year of fierce competition between American technology giants for control of the world's most coveted technology.

Manas, a Singapore-based Chinese-founded company specializing in agent AI for small and medium-sized businesses, announced on Monday that it will join Mark Zuckerberg's Meta, the parent company of Facebook, Instagram and WhatsApp.

Unlike AI chatbots like ChatGPT and Deepseek, both of which require user prompts to perform tasks, Manas claims its product requires far fewer prompts than its competitors and can make decisions and complete tasks on its own.

And unlike many industries that are highly regarded for their potential but are not yet broadly profitable, they do make money, and they do so by selling their products through subscriptions.

The acquisition is intended to be a “bit of a brain transplant” to Meta's existing platform, said Carmi Levy, a London, Ont.-based technology analyst.

Manus' technology could improve Meta's agent functionality, such as answering questions and completing tasks, keeping users on the platform longer and allowing Meta to “earn more from users,” Levy said.

The company will reportedly be sold for $2 billion, a relatively cheap acquisition price in proportion to the potential dividends it will pay to its new owners. The company has been active in AI acquisitions this year to compete with big players like OpenAI and Google.

See | This year's biggest advances in artificial intelligence:

DeepSeek and the biggest AI advances of 2025

CBC's Chris Reyes highlights the biggest advances in AI in 2025 and the continued call to establish clear boundaries to prevent abuse.

Why Meta took action

Although now considered primarily a legacy technology company, California-based Meta is “quickly pivoting to reshape its business for this new era of AI,” Levy said.

“Developing much of that technology in-house is proving difficult because it's not something their culture is built around,” he explained. Instead, tech giants are acquiring smaller companies and integrating emerging technologies into their core businesses “as quickly as possible.”

Back in June, Meta acquired data company Scale AI for more than $14 billion and brought its CEO to the board to help launch a “superintelligence” division focused on in-house AI models, including Llama, an open source large-scale language model.

Gil Luria, an equity analyst at U.S. investment banking firm DA Davidson, told CNBC this week that Meta has poured money into superintelligence and ad tech for merchants and is now trying to force the use of artificial intelligence on consumers through its most popular platforms.

“One of the things they saw with Manus was that it was built into Manus. [Chinese messaging app] WeChat is a model for exactly what they want to do with WhatsApp. With this tool, you can do everything: PayPal, chat, payments,” Luria said.

“So by bringing in Manus and putting him there, Mark Zuckerberg is going to give us the sidekick he's dreamed of, this friend-cum-killing assistant who's going to help us do things,” he said. Luria said this could make it easier for apps to monetize than they currently are.

A man wearing thick-framed black glasses is shown from the shoulders up.
Speaking at an internal event in September, Meta CEO Mark Zuckerberg said he is trying to make his company's AI more reliable through targeted acquisitions. (Carlos Barria/Reuters)

The Facebook co-founder wants Meta to be competitive in consumer AI technology, “not only competing with OpenAI with ChatGPT, but also with Google for distribution through search, YouTube, and all of its other properties,” Luria said.

Chinese roots could trouble U.S. regulators

The deal must first pass through U.S. regulators, which closely monitor Chinese-owned companies because of national security concerns.

The most famous precedent is the dispute between the US government and Beijing-based social media app TikTok. The years-long battle recently ended with parent company ByteDance selling its U.S. operations to a group of American investors.

As with TikTok, the Mehta-Manas agreement “will give pause to those concerned about the Chinese government's access to data collected by these apps and platforms and how that information will be handled,” Levy said.

Some of that tension was on display earlier this year when another U.S. company invested in Manas, which is owned by Beijing-based Butterfly Effect. Venture capital firm Benchmark led a $75 million U.S. funding round for Manus in April, an action that drew criticism from some within the U.S. government.

“Who would think it would be a good idea for American investors to subsidize our biggest enemy in AI? [Chinese Communist Party] Will it use that technology to challenge us economically and militarily? Not me,” wrote Republican Sen. John Cornyn, a member of the Senate Select Committee on Intelligence.

Even if the U.S. government considered TikTok to be an active data hoarder, Levy said the government “hasn't seen anything yet” because Manus has an unlimited capacity to collect vast amounts of information.

“Concerns about data integrity and privacy, and of course geopolitical concerns, will therefore be prominent throughout the regulatory process, and it is not a given that the United States green-lights this agreement.”



Source link