Limitless Podcast - China Blocks $2B Meta's Acquisition of Manus. It's Getting Real

Episode Date: April 28, 2026

 The fallout from Meta's $2 billion acquisition of AI startup Manus, haz beencomplicated by the founders' detention in China and a government order to reverse the deal. We discuss the implications for US-China relations, the rise of China’s independent AI ecosystem, and the strategic challenges facing Meta CEO Mark Zuckerberg.------🌌 LIMITLESS HQ ⬇️NEWSLETTER:    https://limitlessft.substack.com/FOLLOW ON X:   https://x.com/LimitlessFTSPOTIFY:             https://open.spotify.com/show/5oV29YUL8AzzwXkxEXlRMQAPPLE:                 https://podcasts.apple.com/us/podcast/limitless-podcast/id1813210890RSS FEED:           https://limitlessft.substack.com/------TIMESTAMPS0:00 Intro1:13 China's Bold Move3:57 The Stakes for Meta5:15 China's Strategy Shift12:46 China's Energy Advantage13:38 Open Source Models19:22 The Future of Meta and China21:19 AI as a Bargaining Chip------RESOURCESJosh: https://x.com/JoshKaleEjaaz: https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠

At the end of last year, META announced a $2 billion acquisition of an AI startup called Manus. They make AI agents very similar to OpenCore, but they were founded in China. Then in March recently, the two founders visited China on business and were told that they couldn't leave the country. Their passports were taken away from them and they were blacklisted from travel. Then today, China revealed that they have completely banned the acquisition deal by META and have requested that meta reversed the entire deal. Now, this is unprecedented from China. For the last 20 years, they've happily accepted U.S. capital, but this marks the turning point after three years where China has built their own independent AI stack, and they don't need to

rely on America anymore. Yeah, those two people from that meeting last month, that was the CEO and the chief scientist of Manis. They got a summons from the Chinese government body called the National Development and Reform Commission. Basically, it's the agency. that runs the country's economy. They go to the meeting, they get questioned about the meta deal, and then they say, you are not allowed to leave. Now, fast forward to today, four weeks, I will say, before Donald Trump is supposed to land in Beijing to meet with Xi Jinping, China killed the deal. And they did so in one sentence. They said, just undo it. The two men still can't leave. So now not only are the two people stuck in China, but the deal that meta has signed with Manus

is now kind of in a weird gray area. Now, for those who aren't familiar with Manus, Manus is Meta's response to OpenClaw. OpenClaw is the claw. It is the agenic system that kind of runs your life for you. It performs agentic tasks. Long form, it has context, does memory. It was a big deal for meta. And now meta has this problem on their hands where China is saying no. When the reality is there's already an entire Singaporean office setup and the CEO of Manis is on the C-suite of meta already. So there's a lot to unpack here. There's a lot to unwind. The tension between China and the United States has never been higher. I mean, this is pretty unprecedented as it relates to the dynamics between these two countries.

Yeah, this all started around the time last year that Zuck started acquiring or started doing his spending spree. I think he spent in total around $25 billion, $15 billion, which was spent on just one single man. But Manus was part of this spending spree. And the timeline on this is pretty insane. I've got it pulled up here. So in March in 2025, Manus launched the AI agent, right? And it's what you just mentioned. It's like its own AI operating system. And within like 20, 24 hours, or within seven days, rather, they had a 2 million people on the waiting list. And people were like, I remember this fissory, reselling invite codes for like $1,500. $1,500 is crazy.

I actually tried to acquire one because, like, this is before AI agents exist. This is before GPT and Claude had built their own agentic systems. Manus was the first. So the founders, with all the success, was like, oh, we got to like kind of kill the Chinese image so that Americans will use our product. So they ended up moving from Beijing, where they were based, to Singapore, and they raised a massive round from benchmark capital, which is like one of the biggest funds in America.

I think it was like $70 million. And within six months from then, they hit $90 million of revenue run rate. So things are looking amazing. And they thought that they'd subverted Chinese authority simply by moving their company to Singapore, which is obviously not in China. Then in March, 2026, China stops the manor. co-founders from leaving the country, and today they've blocked the entire acquisition. Now,

the main question that pops into my head, Josh, and I don't know if you're able to, like, answer this for me is, are they able to actually do this? I thought the deal was already signed. I thought the ink had already dry. Like, I'm kind of confused how they unwind a deal that's already been dealt. I think this is where there's going to be a problem, right? It's because, like we mentioned, there is an entire Singaporean office where there's over 100 workers who are from Manus working with Meta. The CEO of Manus is, on the C-suite in meta, there's a lot of already kind of baked in progress that has been made after this deal. And China pulling the cord creates a lot of tension. And I'm not sure where this

is going to land because does meta want to actually go to war with China to prevent this blockage from happening? Maybe. This is a huge bet for the company. And a big bet on the future of the company, what this looks like. They were starting to roll this out. It was very deeply integrated. It fell apart. I don't know what's going to happen. This is actually a pretty insane story. Okay, so there are like some crazy parts of this on both sides, right? So for Meta, you've spent $2 billion, you have hundreds of employees from Manus that now are under Zucks employment. Just to be very clear, it's like under the American company, Inc has been sealed, the deal has been done. Now, to your point, Manus is being used a lot in Meta, but specifically, if you haven't used as a user,

it's being used by advertisers. We spoke about this on a previous episode, thing like two weeks, weeks ago, where Manus is being used by advertisers to spin up their advertising projects. And it's actually resulted in a 30% quarterly jump in revenue alone, just by people using Manus. It's cut out a lot of repetitious manual tasks. So it's a very good product. And just because you haven't used it, doesn't mean it isn't valuable to Meadow. They're building it. It's implemented. It's core to their system. Advertising is like a lot of meta's revenue. So now they theoretically have to unwind this. But I think the alarm bell that's ringing for me, Josh, is China's never. It's done this before. They have never unwound a closed deal. I'll repeat that again. They've never unwound

a closed deal that they have signed off on that is all completely legal, which tells me that they are now treating these AI assets, whether you want to call them workers, researchers, the actual product itself, as a kind of national asset, and they're treating it as kind of like a geopolitical asset against the US. They feel like if they don't own this thing, then the US is going to wipe them out. So I started looking at like other parts or trends and I realized China's kind of been like moving in this direction over the last couple of months. Specifically, they have like refused to use Nvidia's GPUs to train their own AI models amongst the top Chinese AI labs. And they've started mandating that all these Chinese labs start building and using the AI models within the China itself and not really open sourcing or open sourcing only parts of it but not the research itself to the US. So there's been this cultural shift where China now doesn't really depend on the US as they had before in the past over the last six months.

They now kind of have their own independent stack. And I think that's probably part of what is playing into this story today. Yeah, man, I kind of feel so bad for Zuck because he's tried this over and over to get this agentic system. He first tried to buy Ilya's safe superintelligence, if you remember. And Ilya said no. And then he went to another open AI co-founder. He went to Miramar, he's thinking machines. She gave him a no.

he reached out to perplexity. They gave him a no. And then they wound up paying $15 billion for half of scale AI just to get Alexander Wang. So they really were struggling to get some help in this agentic world. They finally acquired it through Manus and now they're having this huge problem. And an important caveat that's probably worth mentioning is that it's a little disingenuous. And I could see why China's upset. Because I mean, Manus started off in Beijing. Then they transferred over to Hong Kong. And then they transferred over to Singapore. And this is because, because I mean, American venture capitalists are not going to invest in Chinese companies and China does not want to engage with American acquisitions. So they moved to a neutral place. When they did this, I think 80 Chinese employees were laid off. All of the Webo posts were deleted. All of the new offices were opened up. Everything was a clean slate. So the people who helped them get there, they were cut off. They started fresh in Singapore. But now China's like, wait a second. No, no, no, no. You started here. You are Chinese. You are coming back to us. And what a disaster for Zuck. This poor company, they can't figure it out. They're spending so much money. They haven't been

able to launch a good product. And they finally got Manus, like you mentioned, who was selling these invites for thousands of dollars because it was that good. And now it's ripped out from their hands. So really brutal turn of events. And it's going to be fascinating to follow meta and see how they react to this. Yeah, I think the playbook that Manus founders tried to pull off, but failed at was, oh yeah, if we just wipe ourselves or rather the premise of the actual company, HQ, to another country. We can evade Chinese law, but remember China is more of a dictatorship in this sense and they were like, hey, you can't actually do this. So we can pull your passports and unwind this entire deal because it is still effectively Chinese. It was founded in China.

Now, I just want to comment on China's AI strategy in general, because I think it plays very well into this particular story today. Manus doesn't actually own any AI models. I don't know if If you guys knew that, the people listened to us knew that, but they're kind of like a wrapper over a bunch of different AI models, and typically the AI models that they would use are American AI models, right?

But since, well, actually, over the last seven days, China released three AI models. I think it was Deepseek V4, Kimi K2.6, and just retesting my model knowledge here now, a Quen 3.6. And people have heard about Chinese models all along, right?

They're like, oh, it's open source, but it's never as good. These three models are at Paris, if not as good as Claude Opus 4.7 and GPT 5.4. So pretty much 90% of frontier model capability, or 95% of frontier model capability, you could now access for free, 100% open source, download, run it on your own devices or GPUs if you happen to have a couple laying at home for whatever reason. And they're all Chinese models. Chinese models are actually also being used in a bunch of Silicon Valley startups who can't afford to kind of run and fine-tune their own

models, which of course, Claude and GPT are closed, so they can't actually do that, and it takes hundreds of millions of dollars to train your own models. So the point is, Chinese models are more accessible and more easily fine-tuned, so why wouldn't you use it if you could just access maybe 95% of frontier capability? And the large reason for this is because China has started forcing their AI labs to train on Chinese-made chips. Now, I just want to spend a very quick moment to explain this. Deepseek, moonshot, and... Alibaba have been mandated by the Chinese authorities to not train their AI models on NVIDIA chips. Now, typically, America had banned China from buying the chips. That got reverted,

and America was like, cool, you can have these chips. China said, no, we're good. We're going to train our own chips to be better so that we aren't reliant on America at all. And it seems that Huawei specifically has caught up. That's what all these latest models that I just mentioned are trained on and are inference on. And now it seems like that gap, has close. So it doesn't really matter. This is worrying, from my perspective, at least, because America could keep tabs on China, and now they're kind of given the opportunity to leap progress. They have a heck ton more energy. So this is like the alarm bells ringing from today's story. Like, Manus isn't just about like one startup. It's about China's entire AI strategy.

Yeah. And I mean, everyone knows the name Deepseek now. Deepseek was the ones that released that paper early on. They kind of changed the game as it comes to, I guess, like reasoning in AI models. where OpenA, I kind of figured it out, but Deepseek realized it and then published it publicly and then everyone followed on. Deepseek has always been the frontier of this open source race. And just on Friday, DeepSeek released V4, which is kind of a scary benchmark relative to the other open source models. I mean, I'm looking at the chart here and it actually performs better on a few benchmarks than GPT5.5, which is OpenA's brand new frontier model that just came out last week. And I mean, just last week alone, China dropped three huge models. We got DeepC fee.

which is probably the flagship, but then there was that new Quinn model, the new Kimmy model. They're publishing at a much faster cadence, and you have to imagine that part of this comes from just their boldness and willing to distill the larger models from the United States, but also their capabilities to train these at an increasing rate. I mean, they're clearly innovating on the software side, but something's working on the hardware for them to publish models at this cadence with this type of power. On that note, Gavin Baker put out an amazing analysis. Did you watch the Doha Cash episode where he's like grilling Jensen and Jensen just says like, I'm not a loser like Nvidia's not this? Right? So Gavin Baker commented on this and he said actually

Jensen makes a really good point where the chips that Nvidia builds are specifically built with the understanding that the US has a very limited amount of energy. That's number one. Number two, that a lot of these chips are going to be based on pre-training, which is like the part that comes, I guess, before post-training, obviously. But the point is it's like a large training run, right? It's the bulk of the expense. Now, China took a different approach. They said, well, we're not energy constrained.

We have 3x more energy than the US. And also, we think that building the best AI model happens after pre-training. So in the post-training thing, now one thing that Dario Modi of Anthropic and Sam Omband, of Open AI have announced with their latest models, including Claude Mithos, is inference has been the key to unlock a smarter model. So they're confirming what the Chinese has also confirmed. And so the reason why Chinese chips specifically are doing so well is because they're designed around inference. So the architecture for a Chinese chip, if you were to try and train like a US model, looks very different. And it actually wouldn't work. The hardware actually kind of delineates

at this point. So China, although it seems like we're competing, is building their own kind of like stack themselves that wouldn't necessarily operate or work in the U.S. And I think that's a story that most people aren't addressing at this point. And we're giving China the freedom to go do that right now. Yeah, that perhaps talks to like the increased velocity, right, that they've been having is because they're not spending that much time on pre-training. They're all in on inference, all in on distillation. And we've kind of seen this, I guess, a way that you could track this back to something

that we've seen is how something like GPT 5.0, something nano will actually be far more superior than like the following model, even though it's cheaper and smaller. And it's because it takes advantage of this distillation. And we see this all the time where with Nanobanana, the Google's image generation model, nanobanana, nano, like the small one, was actually better than the pro one, even though it had much less model weights. And I think what we're seeing here is a lot of innovation on that inference side, on just like a faster compute. And Google's been addressing this recently. They split their most recent TPU into two, one just for pre-training one for inference. It's clear that China is very much over-indexing on inference, and right now that's where a lot of the gains are

happening. I just wanted to pull up this post because you just reminded me of something. For those of you who say no one uses Chinese models, the answer is just you're wrong. So Kimmy K-2.6, which was released last week, and it was actually the first model of the three models that we mentioned. DeepCV4 was the last one, is now the number one most used model on OpenRata. OpenRata is this website where you can basically get access to all the models, sometimes before they actually officially release, and it can track token usage. Now, token usage is the metric that you can kind of track to figure out, you know, which model actually is worth its weight in gold, and what are people using it for? OpenRatto reveals that Chinese models like take the number

one spot almost the entire time, and part of that is probably because like all of the closed-sost models don't want to reveal their full metrics. But the point being is the token usage of Chinese models have now surpassed U.S. token consumption, at least from publicly available data. So the data would then indicate that these Chinese models are obviously getting not only better, but more accessible to any and all developers that want to use it. It's not just enterprises that use it. It is just open source developers or just developers that can't afford to train or pre-train their own models. So the Chinese models are way more powerful at this point. They've almost caught up to parity with the US models. I think this probably suggests, if I had to guess,

that the next couple of iterations of Chinese models will probably end up leapfrogging. some of the US ones. This is just a guess from my end. And so it'll probably switch to becoming closed source. I can't see China wanting to open source these things or continue to open source these models if they're doing things like unwinding Manor's deals and trying to keep the assets on mainland China. Well, you know who's probably happy about these open source models doing well? And I'm like, this is a sneaky one, but probably Jensen. And after I watched that interview that he had Dorcas, I think one thing that became clear to me is that open source models, benefit Nvidia more than anybody else.

And I think that's why Nvidia has really been leaning in heavily to be the leader in the United States, at least, in terms of open source AI, because it requires the world to train on an open hardware stack, right? If everyone is creating these models and they're publishing the weights, you can train them on Nvidia GPUs. Closed source models, like someone like Anthropic or OpenAI, who are seeing really heavily sway towards moving to TPUs or accelerators that are not made by NVIDIA, they have the ability to build this whole custom stack built on hardware that isn't NVIDias. And that seems to be like a real

threat to the Nvidia platform. So when Jensen got a little bit heated on the podcast, I have a feeling that has to play at least partially a role, where open source really benefits them. It makes everyone want to use NVIDIA chips. The second these closed source models win, and if everything becomes closed source, there's a strong bias towards these closed loop kind of hardware stacks that Nvidia is no longer in the loop on. And that could create a little bit of an issue for the company. Yeah, I think you hit the nail on the head. In this interview, actually, if those of you haven't watched, definitely go watch it. Jensen says, or was asked, hey, you dominate on GPUs. Why don't you move up the stack and dominate there as well? You obviously know so much about

how this is going to play out. And he answers very simply, I am so hyper-focused on GPs. GPUs and the hardware architecture, and this is the game I want to play, dominate, and own, right? That was his response. So he's hyper-focused on winning GPU architectures and selling the picks and shovels for this entire race. NVIDIA also announced NemoClaw, which is their enterprise version of OpenClaw. Now, you might be asking, okay, why is Nvidia launching that? Why are they launching a bunch of open-source models, which seemingly like no one is using? It's, to your point.

he wants to push adoption of AI in general. Whatever is built on top of his Nvidia GPUs will demand more Nvidia GPUs in the future. That's how he's going to make this money. And guess what happened at the end of last week? Nvidia once again surpassed a $5 trillion market cap. They went above, then they went below, and now they're sitting pretty above it.

So his plan is paying off. He couldn't explicitly say clearly on the podcast that I'm going to make a lot of money if I sell GPUs to China. Obviously, he can't say that. It's in poor taste. It would go against the philosophical argument that he was making, but that is the truth ultimately. Yeah, so the Game of Thrones is getting a little weird now because now we're seeing direct

combat in a way between China and the United States. We're seeing a lot of that happen just domestically across these companies. It is getting higher and higher stakes, and you could see as the stakes could hire, people get a little more emotionally charged, people make larger decisions, and China pulling the rug on this deal sets a pretty crazy precedent. I mean, this has never happened before where they've actually interfered with something on a material scale. Two billion dollars is a lot of money, and it's been months that they have been integrating madness into the platform. Now, Meta, if they do decide to actually unwind this

deal, is kind of sitting in a little bit of an uncomfortable position because they still haven't released their product, and they now don't have a clear path to getting a claw-like operating system for their software that has all this rich data about all its users. So we'll be following the story closely. We'll be following the China story closely. We'll be testing out all of these models, including Deep Seek V4, which I haven't had a chance to play around with, but seems like it's pretty powerful. And I think that's probably the update for China right now. It's just kind of wait and see. Like we have this news that came out. We're going to see how meta responds. We'll see how China responds. And if they're actually willing to unwind this deal for the sake of doing what the Chinese government wants.

Yeah, can we get a response from meta, please? It's been like 12 hours since this, since this announcement. And I'm on the Manus website right now. It says Manus is now a part of Meadow. So they haven't updated it yet. they're probably got a ton of lawyers putting together their response. But if you're listening to this from Meta and you want to give us, you know, undisclos information via an anonymous tip, our DMs are open. Like, feel free to reach out.

But it'll be sad to see something like this go. I think OpenCool started a movement. Manus is probably number two. Maybe Anthropics called Co-Work is number two as well. But it is a good enough product and it gave Meta the upper hand. They were actually making money from this thing. So it's sad to see it happen. I have a question for you, EJS.

Yeah. So Trump is going to China in like two or three weeks to meet with Xi Jinping. So how important of a topic do you think this is? Like is this something that would be raised to that level? Is Trump going to discuss this with Xi Jinping to try to work out a deal? Like, is AI and meta that important that this acquisition go through? I think that's going to be an interesting thing to follow too. Like does this become a topic on the debate table between the two of them?

Okay. So I don't have a tinfall hat, but I'm putting it on, right, to answer that question. Okay, cool. I think China's going to use this as a bargaining chip, pun intended, to get access to bleeding edge Nvidia GPUs. I think this is a chess move, right? Because right now, Nvidia is selling them GPUs, but they're like the old ones. They're the ones like in the warehouse shed, the discarded items, right? They want to like keep China on a leash. I think Xi Jinping is going to go to Trump and say, hey, listen, we can do this ourselves, but we're willing to pay for your GPUs as long as you give us

access to the best one, no frills. And I think he's going to use this as a bargaining chip. Otherwise, we're not going to take any more investment from America. Yeah. That could be. It's going to be an interesting dynamic. We'll see. I love that this has elevated to the global stage. This is now the single hottest topic in the world. So we will be here to follow it. As always, thank you guys so much for watching this episode. I hope you enjoyed. We had an episode that just released yesterday talking to the founder of USVC an interview, a rare interview with the general partner from this fund that gives you access to Anthropics, SpaceX, Open AI, a lot of the large stocks that you would want to be

investing in pre-IPO. That was pretty interesting. And then we'll have a whole bunch of new episodes coming out this week. So if you enjoyed, please don't forget to share with your friends, like, comment, leave a message, give us a five-star review, whatever you want to do. But as always, thank you guys so much for watching. And we will see you guys in the next one. See you guys.