The Data War: China, Meta, and Artificial Intelligence

Y si Meta compraba Manus, una empresa americana tenía acceso a esa tecnología y, posiblemente, a los datos que usó para aprender.

And if Meta bought Manus, an American company would have access to that technology and, possibly, to the data it used to learn.

Para China, eso era inaceptable.

For China, that was unacceptable.

I want to push on this a little.

Because one argument you hear is that China is being protectionist, pure and simple.

Keeping its best assets at home, out of American hands.

But there's another reading, which is that this is genuine security logic, not just economic nationalism.

Los dos son verdad, creo.

Both are true, I think.

China quiere proteger su tecnología porque es un recurso estratégico, como el petróleo o el gas.

China wants to protect its technology because it's a strategic resource, like oil or gas.

Pero también quiere tener las mejores empresas de inteligencia artificial del mundo dentro de sus fronteras.

But it also wants to have the best AI companies in the world within its borders.

The oil analogy is one I hear a lot, and I think it's imperfect but useful.

Oil you can run out of.

Data, in theory, keeps growing.

But the concentration of it, who sits on the largest reserves, that part of the analogy holds.

Y hay otro punto importante.

And there's another important point.

China invierte mucho dinero en inteligencia artificial desde hace muchos años.

China has been investing a lot of money in artificial intelligence for many years.

En 2017, el gobierno chino dijo que quería ser el líder mundial en inteligencia artificial para 2030.

In 2017, the Chinese government said it wanted to be the world leader in artificial intelligence by 2030.

That 2017 announcement was, I think, one of the most consequential policy declarations of the last decade, and it was treated as a minor footnote at the time in most Western newsrooms, mine included.

Y ahora vemos el resultado.

And now we see the result.

China tiene empresas de inteligencia artificial muy fuertes, como DeepSeek, que sorprendió al mundo hace unos meses, y Manus, que también fue una sorpresa.

China has very strong artificial intelligence companies, like DeepSeek, which surprised the world a few months ago, and Manus, which was also a surprise.

No son copias de empresas americanas.

They are not copies of American companies.

Son originales.

They are originals.

DeepSeek was a genuine shock to the system.

When it came out in January 2025 and matched or beat American models at a fraction of the cost, you could almost feel the Silicon Valley anxiety through the screen.

Sí.

Yes.

Y eso explica por qué Meta quería comprar Manus.

And that explains why Meta wanted to buy Manus.

En Estados Unidos, muchas empresas de tecnología tienen miedo de perder la carrera de la inteligencia artificial contra China.

In the United States, many technology companies are afraid of losing the AI race against China.

Meta no quiere quedarse atrás.

Meta doesn't want to fall behind.

There's a TikTok parallel here that I think is worth naming.

The US spent years trying to force ByteDance to sell TikTok to an American company, precisely because it was too powerful and too data-rich to leave in Chinese hands.

China just did the exact same thing in reverse.

Es la misma lógica, sí.

It's the same logic, yes.

Y es un poco irónico, ¿verdad?

And it's a little ironic, isn't it?

Estados Unidos dijo que TikTok era un problema de seguridad nacional porque era una empresa china con datos de americanos.

The United States said TikTok was a national security problem because it was a Chinese company with Americans' data.

Ahora China dice lo mismo de Meta y los datos chinos.

Now China says the same thing about Meta and Chinese data.

The irony is so clean it almost seems designed.

And yet, both positions are internally consistent.

Which tells you something about where we've arrived: two competing systems that have each decided the other cannot be trusted with their most valuable digital infrastructure.

Y esta situación tiene un nombre: desacoplamiento tecnológico.

And this situation has a name: technological decoupling.

Las dos economías más grandes del mundo se separan en el mundo digital.

The two largest economies in the world are separating in the digital world.

Internet ya no es un espacio global.

The internet is no longer a global space.

Hay un internet americano y un internet chino.

There is an American internet and a Chinese internet.

The concept of a 'splinternet' has been floating around in tech policy circles for about fifteen years.

And for most of that time it felt like an exaggeration, a useful provocation more than a description of reality.

It's starting to feel like an understatement.

Y las consecuencias son muy grandes.

And the consequences are very significant.

Si el mundo digital se divide en dos sistemas, las empresas de otros países tienen que elegir: trabajas con tecnología americana, o trabajas con tecnología china.

If the digital world divides into two systems, companies in other countries have to choose: you work with American technology, or you work with Chinese technology.

No puedes usar las dos fácilmente.

You can't easily use both.

Which puts places like India, Brazil, Southeast Asia, and frankly Europe in an uncomfortable position.

Do you pick a side?

Do you try to play both?

And what do you lose either way?

Europa quiere tener su propia tecnología de inteligencia artificial, independiente de Estados Unidos y de China.

Europe wants to have its own artificial intelligence technology, independent of the United States and China.

Pero todavía no tiene empresas tan grandes como Meta o como las empresas chinas.

But it still doesn't have companies as large as Meta or the Chinese companies.

Es un problema serio para Europa.

It's a serious problem for Europe.

And the speed of this matters.

AI development right now is moving so fast that falling behind for even two or three years could mean you're not just a step behind, you're in a different race entirely.

Sí.

Yes.

Y hay una pregunta muy difícil detrás de todo esto: ¿quién controla la inteligencia artificial controla el futuro?

And there's a very difficult question behind all of this: does whoever controls artificial intelligence control the future?

Yo creo que sí.

I think so.

Y por eso este bloqueo de China es tan importante.

And that's why this Chinese block is so important.

I spent time covering conflicts where the question was who controls the oil, who controls the ports, who controls the roads.

This feels like the same argument dressed in different clothes.

The resource changed.

The logic didn't.

Y lo preocupante es que esta guerra tecnológica no tiene reglas claras.

And what's worrying is that this tech war has no clear rules.

En la guerra comercial, hay la Organización Mundial del Comercio.

In trade war, there's the World Trade Organization.

Pero en la guerra de la inteligencia artificial, no hay una institución internacional que ponga límites.

But in the AI war, there's no international institution setting limits.

That's the part that genuinely unnerves me.

We built frameworks, imperfect ones, for managing nuclear weapons, for trade disputes, for financial crises.

We have nothing comparable for AI yet, and the technology is not waiting for us to figure it out.

Completamente de acuerdo.

Completely agree.

Y mientras los gobiernos discuten, las empresas ya trabajan.

And while governments argue, companies are already working.

La inteligencia artificial ya cambia la medicina, la educación, la economía.

Artificial intelligence is already changing medicine, education, the economy.

El bloqueo de Manus es solo un momento pequeño en un cambio muy grande.

The Manus block is just a small moment in a very large change.

All right, one thing I noticed during this conversation, and you used it a few times naturally, was the phrase 'hacerse con.' You said Meta quería hacerse con esta tecnología.

That's not something you'd find in a basic dictionary entry for 'get' or 'acquire,' but it's everywhere in natural Spanish.

Sí, 'hacerse con algo' significa obtener algo, tomar control de algo.

Yes, 'hacerse con algo' means to obtain something, to take control of something.

Es muy común.

It's very common.

Por ejemplo: 'China se hizo con la medalla de oro,' o 'el cartel se hizo con el control de la ciudad.'

For example: 'China took the gold medal,' or 'the cartel took control of the city.'

So it carries a sense of gaining possession, almost seizing something.

It's more active than 'tener,' more purposeful.

Like there was effort involved, or competition.

Exacto.

Exactly.

'Tener' es solo poseer.

'Tener' is just to possess.

'Hacerse con' implica un proceso, una acción.

'Hacerse con' implies a process, an action.

Alguien quería algo, trabajó para conseguirlo, y al final lo obtuvo.

Someone wanted something, worked to get it, and finally obtained it.

Es la diferencia entre 'tengo un café' y 'me hice con el último café de la mañana.' [chuckle]

It's the difference between 'I have a coffee' and 'I got my hands on the last coffee of the morning.'

That actually captures this whole story perfectly.

Meta tried to hacerse con Manus.

China decided they couldn't.

And now everybody's trying to hacerse con the future of AI.

I'm going to use that phrase all week and probably destroy it somehow.

Seguramente sí.

Most likely yes.

Pero al menos no vas a decir que estás embarazado esta vez.

But at least you won't say you're pregnant this time.