On April 27, the deal between Manus and Meta, which had been rumored for months, finally came to fruition.

It's not "continued review".

It is not "supplementary material".

Instead, they prohibited investment and demanded the cancellation of the transaction.

These words are very heavy.

Because it expresses not just a business opinion, but an attitude.

Buying one less company is certainly not the end of the world for Meta.

Manus losing $2 billion is not the first time the AI industry bubble has burst.

In the startup world, stories of failed fundraising, failed mergers and acquisitions, and zero valuations are nothing new.

But this time it's different.

In the past, many entrepreneurs were used to viewing their companies as purely commercial entities.

A good product leads to user growth, increased valuation, and capital exit – this is a very natural path.

But AI is not the same as the previous generation of internet businesses.

AI is not a new app category, nor is it a smarter office tool.

AI is becoming the key to winning the next round of competition.

Whoever controls the modeling capabilities controls the next generation of software access.

Whoever controls intelligent agent products is likely to control the next generation of workflows.

Whoever controls the AI infrastructure and application ecosystem will have an extra advantage in the future division of labor in the industry.

It wasn't just one company that stumbled.

It's an old world model that's broken.

PART.01, The Old World Model is Invalid

Over the past decade or so, Chinese entrepreneurs have actually developed a very mature, default script in their minds.

People are in China, the market is in China, engineers are in China, and products are grown in China.

But financing can be done in US dollars, the legal entity can be based in the Cayman Islands, the listing can be done in the United States, and if necessary, the office can be moved to Hong Kong, Singapore, or Silicon Valley.

This system has been running for many years.

The underlying premise is:

China needs growth, the US needs assets, capital needs exits, and entrepreneurs need stories.

If everyone cooperates in a gray area, and can grow the company, allow investors to exit, and help the founders get out of trouble, then it is considered a success.

The core contradiction of that era was not "who owns the technology," but rather "how to make the company grow, how to exit capital, and how to continue growth."

As long as this overarching logic remains, many ambiguities can be tolerated.

You can have a Chinese business, US dollar capital, an offshore structure, or a US listing.

Everyone knows there are many gray areas here, but that era was willing to leave an outlet for these gray areas.

But AI is different.

AI is not group buying, food delivery, e-commerce, or an upgrade to short video recommendation algorithms.

AI has now been placed within a competitive framework.

Model capabilities, engineering talent, training data, inference systems, intelligent agent products, and commercialization entry points—any one of these can be considered a strategic asset.

If you then try to apply the worldview of 2010s internet companies to deal with AI companies in 2026, you will run into problems.

Many people don't have a problem with their judgment, but rather that their world model hasn't been updated.

They thought it was a game of capital, but the tables had already been turned.

Previously, you mainly dealt with investors, users, exchanges, and M&A lawyers.

Now you also have to deal with security reviews, export controls, technology boundaries, and competition.

This is not just a change happening in China.

The United States is also changing.

In the past, global capital believed in efficiency.

Capital flows to where there is cheaper talent;

The company will go wherever there is a bigger market.

The project will go public wherever the valuation is higher.

The underlying logic of globalization is the efficiency of resource allocation.

But today, the underlying logic of globalization is becoming a boundary.

Technology has its limits.

Data has boundaries.

Computing power has its limits.

Capital has its limits.

There are also beginning to be boundaries for the movement of talent.

This is what many entrepreneurs find hardest to accept.

It's not that I stopped working hard, or that the product had no future; it's that the mental map in my head had become outdated.

What entrepreneurs fear most is not hard work.

Entrepreneurs all know how hard it is.

What entrepreneurs fear most is running forward with all their might, only to find halfway through that the map has changed.

PART.02 Location, location, and more location

In the business world, many people like to talk about ability.

Product capabilities, financing capabilities, growth capabilities, organizational capabilities, technological capabilities, and narrative capabilities.

These are all important, of course.

But in a great era, the most important thing is often not ability, but position.

Where are you standing?

To whom do you prove your worth?

From whose soil did you grow up?

To whom do you entrust your technological assets at critical moments?

These questions usually seem very abstract.

Entrepreneurs don't like to include this in their fundraising presentations, and investors aren't necessarily keen to ask about it either.

People are more concerned about ARR, DAU, retention, valuation, and who will invest in the next round.

That's how the business world works; when there are too many metrics, it's easy to pretend you're clear-headed.

But once the documents are finalized, these issues become anything but abstract.

The problem with Manus is not that it wants to go international.

Chinese companies can certainly go international.

It's not that it wants to earn US dollars.

It's not a sin for a startup to want to exit.

The real problem is that it doesn't seem to have figured out its own position.

If you decide from day one to be a completely American company, then you should register, raise funds, hire, conduct research and development, serve customers, and accept local regulations in the United States from day one.

If you decide from day one to be a fully Singaporean company, then you should also build your team, conduct business, establish compliance, and accept local rules in Singapore from day one.

These are all choices.

But if you grew up in China's technology ecosystem, benefited from the Chinese engineer dividend, and enjoyed the attention given to China's AI startup narrative.

If a company establishes early influence in the Chinese market and on the Chinese internet, and then, during a period of soaring valuations, uses structural design to repackage itself as a "non-Chinese asset," and ultimately sells it to an American giant, problems will arise.

Because in this era, identity is not something you declare yourself.

Your identity is determined by your history.

Where you accumulate technical skills, recruit core talent, gain initial attention, validate your product, and build team capabilities—all of these things will become part of your identity.

Entrepreneurs can change their registered address.

You can change offices.

The financing entity can be changed.

You can change the PR diameter.

But it's difficult to change one's own life history.

This is the coldest part of the Manus incident.

It doesn't just ask you where you are now.

It's still asking you where you're from.

PART.03 It Must Be Valuable

In China, many people are reluctant to talk about risks when starting a tech business, which is understandable.

Entrepreneurs like to talk about products, users, cash flow, and the next round of financing.

The others sound too distant, too heavy, and too uncomfortable.

But not talking about it doesn't mean it doesn't exist.

As long as the grand narrative hasn't yet impacted your fundraising, exits, compliance, mergers and acquisitions, and team turnover, you can certainly pretend you're just an ordinary entrepreneur.

But once AI is brought into the competition, even if you don't talk about it, it will talk about you.

The value of a technology company within a system can be roughly categorized into three states.

The first type is positive value.

You can fill technological gaps, enhance industrial capabilities, and give China more leverage in a key area.

You don't necessarily have to shout slogans, but your presence objectively strengthens the system.

The second type is zero value.

You're just an ordinary commercial company, making your own products and earning your own money. It's not that important or dangerous.

The system may not care about you, and you may not have any impact on the overall situation.

The third type is negative value.

You may have grown up within China's technological system, but in the end, you might end up handing over your key teams, technological assets, product experience, and strategic narratives to an American giant.

At this stage, the question is no longer whether there is a contribution, but whether it will create a demonstration effect.

If you become a negative value, you are easily made into a typical example.

This statement may sound harsh, but it reflects a critical assessment.

It's not because any particular company is especially important, but because it represents a path.

What the system needs to eliminate is often not a single company, but the path itself.

If Manus succeeds, what will happen next?

A group of Chinese AI entrepreneurs will see this path.

First, establish the technology and influence in China.

Then move to Singapore.

They were then sold to major American manufacturers.

Then use "global entrepreneurship" to explain everything.

If this path proves successful, it will set a very bad example for China's AI ecosystem.

Because it will tell those who come after it:

You can first reap the benefits of China's technology ecosystem, and then, once you've grown large enough, cut yourself out, with American tech giants ultimately acquiring and integrating you.

From an entrepreneur's perspective, this is certainly very tempting.

But from a competitive perspective, that's a different story.

So the suspension of Manus trading is not the main issue.

The key point is to tell those who come after:

This path is blocked.

PART.04 You can't have it all, want it all, and want it all again.

Therefore, this does not mean that all entrepreneurs must stay, nor does it mean that all companies with Chinese backgrounds cannot go overseas.

On the contrary.

AI startups should be globalized.

Good products should serve users worldwide.

Chinese entrepreneurs should not confine themselves to a single market.

Tools, agents, content production, enterprise automation, developer services—these things naturally have a global market.

But going to sea and cutting off are not the same thing.

Globalization and arbitrage are not the same thing.

You have a choice, but don't fantasize about not paying the price.

If you don't want to be part of China's technology system, you should have left on day one.

Go to the US, go to Singapore, go to Europe, it's all fine.

From day one, we were there to get money, recruit people, build a company, develop products, and accept local regulations.

This is a clear choice.

You can also stay.

Acknowledge that you are part of China's technology system, serving China's industries, participating in China's AI ecosystem, accepting the rules, and clearly understanding the relationship between your capital, market, technology, and system.

This is also a clear choice.

The real danger lies in the middle of the road.

We need to start at Chinese speed.

When seeking funding, a valuation in US dollars is required.

It needs a global halo when it is disseminated.

When exiting the market, a major American company needs to take over.

The system arrived, and then claimed to be just an ordinary commercial company.

This is not globalization.

This is an incomplete selection.

In the past, many people called this state of being smart.

Because in a period of peace, ambiguity is a kind of space.

You can maneuver between different rules, switch between different markets, and seek maximum benefit within different narratives.

However, ambiguity is a risk during a conflict cycle.

When the technological competition between China and the United States enters the AI level, many things that could previously be blurred will be redefined.

You used to say that capital has no borders.

Now people will ask, who has benefited from capital?

You used to say that technology has no borders.

Now people ask, where does technology ultimately go?

Previously, you could say that startups were simply business entities.

Now people are asking, which side will this company be on in the next round of industry competition?

This isn't a question only China is asking.

The United States is also asking this question.

So it's not just China that has become stricter; the whole world has changed.

Previously, global capital believed in efficiency; now, national competition believes in boundaries.

If entrepreneurs fail to recognize this change and continue to use the overseas expansion models of the previous generation of internet companies to deal with AI companies, that's not courage, it's a misjudgment.

PART.05 Run forward, don't look back

So the lesson this teaches all AI entrepreneurs is actually quite simple:

Run forward, don't look back.

This isn't just empty encouragement.

I'm not telling you to rush in blindly.

It means that if you choose a path, you must accept the cost of that path.

If you're going to run an American company, then run an American company from day one.

If you're going to start a company in Singapore, start as a company in Singapore from day one.

If you want to run a Chinese company, then acknowledge that you are part of the Chinese system and think clearly about your relationship with technology, capital, market, and regulation.

Every position has its advantages.

Every position comes at a price.

The most common mistake entrepreneurs make is wanting the benefits of each position without wanting to bear the costs associated with it.

If you choose China, don't fantasize about designing your own path entirely according to the exit logic of the previous generation of dollar-denominated internet companies.

If you choose the United States, don't fantasize that you can always access China's technology ecosystem at low cost.

If you choose Singapore, don't assume it's just a neutral shell that can erase all your historical origins.

The most dangerous people are those who never want to make a choice.

This is an era that requires a greater understanding of one's position.

You need to know who you are.

You need to know where you come from.

You need to know who you want to prove your worth to.

You also need to know which side you're on when the great era begins to redraw its course.

This is not a moral performance.

This is a strategic judgment.

Many entrepreneurs like to say that they only care about the product, the users, and the cash flow.

That's certainly true.

However, if the changes of an era can determine your financing, exit, mergers and acquisitions, compliance and company identity, then recognizing the era itself is also a kind of entrepreneurial ability.

It may even be the most important entrepreneurial ability.

Because if the product is made incorrectly, it can be corrected.

If the timing of financing is wrong, it can be corrected.

But if the position is wrong, the cost will be very high.

In the face of a great era, intelligence is no longer the most important thing.

Location is what matters.

Run forward, don't look back.

It's not because the area ahead is necessarily safe.

It's because the path back is no longer there.

Meta plans to lay off approximately 10% of its workforce, affecting 8,000 positions, starting May 20.

Popular Articles