Manus AI's ‘de-China' playbook is a trap

The Star

3 days ago

WHEN Chinese startup Manus previewed an artificial intelligence (AI) agent earlier this year, it went mega-viral. It came on the heels of DeepSeek, when global excitement over China's AI breakthroughs was at a fever pitch, and nobody wanted to miss out on the next surprise hit.
Now, Manus is doing everything it can to sever any ties to the mainland. It relocated its headquarters to Singapore, and its three co-founders have made the move abroad as well.
Butterfly Effect, the company behind Manus, reportedly eliminated all its China-based jobs last week.
It has also scrubbed content from domestic social media platforms Weibo and Xiaohongshu (also known as RedNote), despite maintaining an active presence on X.
Users in China trying to access the site this week were met with the message that it's 'not available in your region,' a departure from a previous memo stating that the Chinese version was under development.
It's the choice these tech firms are forced to make in the current geopolitical climate: stay in the hyper-competitive domestic market or go for more lucrative growth overseas. You can't have both.
Moving abroad means going through the arduous process of trying to 'de-China' the company's origins.
It's a shame given this background is what drove so much hype about Manus in the first place.
Chinese firms have also traditionally had an edge in consumer tech, with access to a vast pool of affordable engineering talent and a hardworking culture. Plus, dubious identity rebrands almost never work.
Manus' decision to recast as a Singapore company follows the furore that emerged in the United States after prominent Silicon Valley venture capitalist (VC) firm Benchmark (an early backer of the likes of eBay Inc and Uber Technologies Inc) announced it was leading a US$75mil funding round in Butterfly Effect.
Fellow VCs accused Benchmark of 'investing in your enemy' and equated it to backing Russian efforts during the space race.
The plans have also come under US Treasury Department scrutiny over new rules related to investments in certain Chinese technology.
It will be very hard for Manus to ever rid the China label from its story, especially after all the attention it received.
Co-founder Ji Yichao was on the cover of Forbes China more than a decade ago, as one of the 30 under 30 entrepreneurs in the country. State-backed mouthpieces have also celebrated Manus's rise, so trying to cleanse its Chinese-ness risks domestic backlash.
It's not the first time this has happened. ByteDance Ltd's TikTok has gone to great pains to rebrand as an American and Singaporean company and assuage Washington's fears about its Beijing origins. But none of this stopped the United States from passing a law last year requiring the parent company to divest from the app that doesn't even operate in the mainland, or be banned due to perceived national security concerns. This doesn't bode well for Manus.
The AI sector has become a lightning rod for China hawks in the United States, who view any consumer-facing application using the tech from its geopolitical nemesis as a threat.
AI video startup HeyGen Inc garnered investment and a raft of US customers after making the move from China, yet was still singled out by lawmakers over potential ties to the Communist Party.
Its co-founder said it's been 'disappointing' to see his heritage treated as something he should 'be ashamed of.'
Wrapped up in national security worries is more than a hint of xenophobia.
Targeting consumer tech firms based on the national origins of their founders is an ineffective strategy.
US concerns about potential threats that data could leak to China or that the CCP could influence algorithms should implement more comprehensive rules to mitigate these risks.
When it comes to buzzy new technology like the Manus AI agent, industry-wide standards are overdue.
Tools that are designed to allow software to take on increasingly complex tasks on their own carry unique risks, including who is liable when things go awry and how much control should be ceded to machines.
Global policymakers should address these concerns regardless of where the AI agent comes from.
Icing out the best and brightest tech minds risks leaving Silicon Valley blind to innovation happening elsewhere. It's in America's interest to do more to support these founders by bringing their talents and breakthroughs to the United States. —Bloomberg
Catherine Thorbecke is a Bloomberg Opinion columnist covering Asia tech. The views expressed here are the writer's own.