Huai Wang: There Are Only Two Doors Leading to the AI Future, One Is in the US and the Other Is in China | End of 2025

"The adaptable survive; the evolved lead." By Harry Wang, Linear Capital

Harry Wang has always carried himself with an engineer's rationality and alertness.

Ten years ago, as one of the most influential Chinese engineers in Facebook's early days, when he dissected Silicon Valley methodologies in his bestseller The Facebook Way, AI was still a sci-fi term that only "crazies" would discuss. Today, with a technologist's romantic streak, he has transformed Shakespeare's famous question into an era-defining challenge: "To be or not to be: in the AI-dominated future."

At the crossroads of the so-called "AI era," everyone has the right to make their own choice. But as a VC investor standing on the front lines, Wang hasn't turned himself into a flag-waving evangelist. Instead, he has made himself and Linear Capital into an experimental field witnessing dramatic restructuring. He requires everyone on his team to use the best AI models from both China and the US, with full reimbursement. Not as a perk, but to keep breathing during "survival training." He has them ask of every new problem: "Can AI do this for us? Can I complete this with AI's help?" This has fundamentally changed a VC's workflow.

"Those who don't embrace AI will die." In 2014, he named his investment venture "Linear." Faced with AI's non-linear acceleration today, he has become the one leading the charge to break inertia. And all of this points to the sense of crisis and excitement within Wang.

In the process of AI reshaping everything, the experience, methodologies, and even aesthetic sensibilities that this generation of investors has believed in for the past decade now face the possibility of complete liquidation. For many investors, including Wang, this is no longer a game about "returns" — it is a contest for "admission tickets." If you cannot become an AI "native" at the first opportunity and lead the direction of the era, then no matter how accomplished your past, you risk becoming a "digital refugee" of the new age in an instant.

Zoom out further: when AI begins to take over logic, replace creativity, and even attempt to simulate emotion, humanity's superiority as the "spirit of all things" is being stripped away layer by layer. We stand collectively before a vast void: if thinking can be computed, if judgment can be predicted, then to what corner should human value retreat?

This "End of 2025" piece is Harry Wang's answer, published on the final day of 2025. And to every reader of Waves, happy new year!

"End of 2025" is still accepting submissions. We welcome thinkers from the business world and capital markets to reach out via our backend or chenzhiyan@36kr.com. Believe that in changing times, only true thinking endures.

Part 01

"50% Chinese by Volume" America vs. "100% Chinese by Volume" China

We were talking about AI ten years ago. Back then, if you said AI would become a monopolistic core force of the future, everyone would have thought you were crazy — because neither technologically nor commercially could that claim be proven.

But today, whether you believe in AI or not, it genuinely concerns the survival of us as individuals and as a species. The speed at which this world changes always exceeds even the most optimistic expectations.

I come from an engineering background; I studied machine learning in my PhD. But this new wave of AI capabilities, from ChatGPT 3.5's launch to now, has continuously refreshed my expectations of its upper limits. What's stunning sometimes isn't the speed itself, but the acceleration.

I believe that before this superpower of AI, God has opened only two doors on earth: one in America, one in China.

The first door is in America. I must acknowledge Americans' pioneering spirit in this domain — sometimes quite savage, but capable of achieving zero-to-one. They have wild, sky's-the-limit ideas. They dare to think what no one dares to think, dare to attempt what no one dares to attempt. Sometimes this temperament may make people uncomfortable, but you cannot deny its effectiveness.

The second door is in China. Chinese people have unique advantages — our one-to-hundred scaling capability has virtually no match worldwide. We have patience, strong vision, and more importantly, a persistence in "taking things to the extreme." Chinese people have historically been very pioneering, but after reaching a certain point, we would content ourselves with Greater China and be unwilling to go further — which of course was fortunate for other regions at the time.

But now, the times have changed.

These two doors, in my view, are the doors to future survival. How to choose between them? As an individual, you may only be able to go deep through one door; but as an organization or institution like ours, you must simultaneously understand the rules behind both doors.

Precisely because of this, we did one thing internally at Linear: we require all our teams to subscribe to and use the best AI models from China and America, with full company reimbursement. The amount may not be much, and it's hardly a benefit — but it is survival training for the AI era.

Having employees simultaneously use ChatGPT, Gemini, and Claude, as well as DeepSeek, Kimi, and others, reveals an interesting phenomenon: American models do have advantages in data openness and creative divergence, but often fall into certain forms of "self-censorship"; Chinese models are actually more open on dimensions like DEI. So depending on what you're doing, if you want a relatively complete perspective, you should reference the best models from both China and America.

In the AI era, whose model you use determines your angle on the world — a single perspective is fatal.

Speaking as a businessman, we haven't invested in foundation models ourselves. The main reason is that as an early-stage fund, putting money in often amounts to tossing it into a bottomless pit — the reputation might be nice, but the returns may not be that substantial. China's best foundation models have taken a different, more cost-effective path toward inclusive AI. On this point, I've prepared myself to be wrong. If I get slapped in the face, I'd actually be happy.

In the China-US AI competition, a long-neglected variable is emerging. Referencing alcohol content, I call it CBV — "Chinese by Volume." I've done some data research; it may not be perfectly precise. Among top American AI researchers, the Chinese proportion approaches 50%; at OpenAI, the proportion of Chinese researchers is also very high. This means that at the talent level, the current AI competition is essentially a contest between a "50% Chinese by Volume" America and a "100% Chinese by Volume" China. When I was an engineer at Facebook over a decade ago, there was a similar situation, but back then it was 25% vs. 100%.

The underlying logic is the real distribution of talent and engineering capability. Behind AI lies mathematics and engineering — tedious matrix operations and parameter tuning. And Chinese people have unique advantages in patience, discipline, and scale on these two fronts.

I have a half-joking formulation: AI model competition is a contest between Chinese people in America and Chinese people in China. But there's a key difference: top Chinese talent in America is very strong in innovative breakthroughs, while本土 talent in China may have more advantages in engineering, scaling, and continuous optimization. These are two different capabilities, both of which will be crucial going forward.

Part 02

Re-examining Human Value

Let me share my views on the future. Speaking too far out is meaningless; I can only speak within our lifetimes. I have a judgment that may sound brutal but that I firmly believe: within the next one to three decades, AI will certainly achieve dominance; by 2030, 30% of existing human jobs will be replaced by AI; by 2040, another 40%; by 2050, humanity's existing work system will be fundamentally restructured — there will basically be nothing left for humans to do.

Many people find this timeline too aggressive. But honestly, I may still be conservative here. Why? Because the concept of acceleration is crucial here, as well as paradigm shifts or paradigm leaps. We can hardly understand this change through linear logic — although our firm is called Linear, we have always encouraged non-linear thinking to understand this change.

Let me give a simple example: mobile phones were commercially deployed in the 1980s, yet by the early 1990s penetration was still below 1%. But by 2000, penetration had begun to surge. Today, apart from some remote areas, global mobile phone penetration exceeds 100%. AI's penetration curve will only be steeper — ten times steeper, even a hundred times.

Of course, what's most unsettling isn't the speed of replacement but the logic of replacement. Many people assume their work is creative and unique, but actually it may not be — or at least, most of it isn't. The brutal truth may be: your "flash of inspiration" has probably occurred many times before in the long river of human history.

And AI's power lies in its ability to learn from similar "flashes of inspiration" recorded by someone somewhere in mathematics, physics, chemistry, or any other field a hundred or two hundred years ago, and ultimately deliver that "flash of inspiration" into your content.

When creativity becomes a scalable "commodity," where is our value anchor? I'm not preaching pessimism. On the contrary, I think this gives us an opportunity to reconsider "human value." If repetitive creative work can be replaced, then what is truly human and irreplaceable?

Of course, everyone knows that current large models have a problem — they "speak nonsense with a straight face." This is the original sin of probabilistic models: the choice of the next token is often based on infinitesimal probability differences. The best answer and the second-best, third-best answers may differ by only 0.0001% in probability.

But I believe that to a large extent, AI is already "eating from our bowl." This imperfection reveals a more brutal truth: AI doesn't need to become superhuman, doesn't need to score 100 on every subject to replace you — it only needs to exceed expectations for your specific position across comprehensive dimensions.

Many people are still waiting for the ultimate form of AGI (Artificial General Intelligence) or ASI (Artificial Superintelligence), refusing to believe that day will come soon. I still believe the vast majority of jobs will be replaced, so we proposed a concept internally: "Everything can be AI" — not as a slogan, but as our daily behavioral guideline. When you encounter a new problem, first instinctively ask "Can AI do this for us? Can I complete this with AI's help?" This has completely changed our workflow.

This sounds simple, but is extremely difficult in practice. Because for most of us, all the experience, intuition, and workflows formed over the past two to three decades were built on inertial logic from the "Pre-AI era." Whether you can enter the next era depends on whether you can personally break these inertias and place AI at the premise of all thinking.

I have substantial daily reading needs. Now, I first use Notebook LM to digest books and generate podcast scripts. I walk two to three kilometers to and from work every morning and evening — perfect time to listen to these AI-generated podcasts. AI doesn't necessarily deepen my understanding of problems, but greatly boosts my efficiency; then I decide which content merits deep reading. Now, when approaching new content without AI assistance, it's not that you can't read well — but the time required to reach the same level of understanding would be far greater than others'. In an era where speed determines everything, efficiency gaps are gaps between life and death.

This even creates a new form of "social filtering": previously we might have focused on a person's values, professional background; now we may also need to look at how they use AI, how they view AI. We put forward a rather direct view: "Those who don't embrace AI will die" — this death isn't physical death, but the death of your competitiveness, the death of your value of existence.

In my view, no matter how brilliant you were in that generation, those who cannot learn and recognize "AI-lize everything," "AI first" will be eliminated. Many renowned successful entrepreneurs and investors — maintaining the status quo, your lives may be fine. But if you want to continue succeeding in the new era, you must become "AI-first" people, and collaborate more with younger generations, understanding their methods and ways of thinking.

The sunset of an old era and the rise of a new era — it's a long tail, not something that changes overnight. If you cannot embrace these new changes, you're not losing to machines; you're losing to people who can use machines.

Part 03

Beer Foam and Soap Bubbles Are Different Kinds of Foam

Many people ask me: China's AI field is so hot right now, investment and financing so frenzied — is this a bubble?

I answered this question seven or eight years ago. My answer was: There are two kinds of foam. One is soap bubbles — when they pop, nothing remains, just a mist on the ground. The other is beer foam — beer with foam tastes much better than beer without, and beneath the foam lies substantial, aromatic beer.

On the battlefield of computing power, Nvidia remains the undisputed king. Jensen Huang is not only a great engineer but also a great salesman — he knows the right thing to say and do at the right time. In my view, Nvidia's greatest pressure is China's rise in computing power, but this still needs time.

Perhaps there is foam in AI now, but we still believe that in many domains it may be beer foam. In embodied robotics, for instance, the "ChatGPT 3.5 moment" hasn't yet arrived. When it does, I believe this may be the greatest revolution since the Industrial Revolution — and American anxiety will rise another notch. So although we've always felt there's foam, we've still resolutely deployed substantial capital.

In tech investing, you must respect the present state while also being able to think about the future for them. When things truly land, applications will present a massive J-curve. So in early stages, investors need to maintain steady mindsets. Regardless of short-term market noise, in the long term this direction is the inevitable path of physical-world digitization.

When ChatGPT 3.5-level capabilities are injected into robots, it will be the greatest transformation since the Industrial Revolution — because it will no longer just process digital signals, but directly manipulate atoms.

What will future factories look like? I can describe a scenario: various robots in a factory, sharing the same base architecture but with quickly swappable end effectors. They can interact with each other — possibly in language we understand, or possibly not, as long as they're given a clear goal they can autonomously operate. Which machine plays what role mainly depends on reconfiguring perception modules. Today these robots produce phones; tomorrow they may switch to drones — just a new model pushed from the cloud, adjust the end effector, and the entire conversion may take only minutes.

How to more efficiently and cost-effectively realize this vision? In China, I see only two places: the Yangtze River Delta and the Pearl River Delta. Behind this is the spillover effect from China's electric vehicle supply chain in these regions. The future of robotics is essentially the second explosion of China's EV industry chain. These supply chains highly overlap with automotive, especially electric vehicle supply chains. In a sense, the future of robotics is an extension of the EV supply chain.

In recent years, China's rise in consumer electronics has also been evident. China has produced or OEM'd too many electronic devices for the world, building powerful supply chain capabilities in this space, and in the process cultivating many talents who don't just understand technology, products, and supply chains — more and more companies understand how to do distribution and marketing.

This evolution of capabilities is giving birth to a new generation of entrepreneurs. They don't just understand technology; they understand human hearts. So we believe that going forward, there will be many entrepreneurs who both know how to make a good product and know how to sell a good product. When AI algorithms can truly understand emotions and are cleverly packaged in hardware you can't put down, things change. It's no longer a tool; it becomes part of your emotions — you'll want to use it, you'll depend on it. This is the decisive factor for consumer electronics in the next decade: who can make products that people "want," not just "need."

Human beings are becoming increasingly lonely, and you are creating value.

We also pay close attention to AI's acceleration of the journey from scientific discovery to commercialization — new materials, chemistry, biomedicine, and so on. These are all important directions. Today, whether in mindset or organizational structure, research and industry are often integrated. In my view, future IPOs and Nobel Prizes may well belong to people at the same company. This would have been hard to imagine previously, but is entirely possible in the future.

In this situation, we truly hope not only Chinese enterprises benefit, but that the entire world can benefit from technology. Because science itself really shouldn't have national borders behind it, though commercialization does need national borders. This returns to Linear's enduring vision: "Frontier Tech, Frontier Productivity, Frontier Lifestyle — All for a Better Society."

Part 04

Globalization Isn't Dead, It's Just Restructuring

Finally, I'd like to briefly discuss geopolitics. I've referenced scholars who study historical mega-cycles. Historically, the cycle of global dominant power transitions is roughly 150 years. Portugal, Spain, the Netherlands, Britain, America — each leader dominated for about 150 years. Counting from when America gained global leadership after WWII, nearly 80 years have passed. By this calculation, the next 30-70 years may be a window of power transition.

But this won't be a blitzkrieg; it will be a game of "tai chi." Americans excel at Western cowboy-style confrontations — open guns, open knives, face-to-face combat; while China is using strategies of "when the enemy advances, we retreat; when the enemy retreats, we advance," playing tai chi, borrowing and neutralizing force to stretch confrontation into a prolonged game.

Our investment strategy has also shifted somewhat: previously we mainly invested in China domestic projects; now we pay more attention to overseas teams that can integrate global resources, particularly those that can effectively leverage China's supply chain.

Globalization isn't dead; it's just restructuring. In the next decade, only those who understand, value, and seize the new opportunities brought by AI development and geopolitical transformation can create tremendous value.

In this era of dual shock from AI and geopolitics, survival itself is the most profound innovation. We no longer pursue "predicting the future" — that would be too arrogant — but even so, we hope everyone can benefit. What we've learned is to continuously adapt amid uncertainty: be adaptive!

I'd also like to share our new standard at Linear Capital for finding founders. In recent years, we've found that excellent founders need a particular trait: "healthy paranoia." They truly believe they can change the world, but are willing to start by changing themselves. In good times they remain vigilant; in bad times they remain optimistic.

I've seen too many founders who want to change the world but are unwilling to change themselves. And those who truly accomplish things are always those who first change themselves, then influence others.

Our generation of tech investors is fortunate — we've witnessed American technological innovation and personally experienced China's scaling rise. We respect America, but we are also Chinese. This identity can sometimes feel撕裂, but we strive to build bridges between the two worlds.

An interesting phenomenon: this year, more entrepreneurs, founders, and young people are returning to the market, and more talent is returning from America. But I must also say: the easy money is gone; the hard money is the battlefield for heroes. In this era, what matters is whether you've already set out, whether you've already broken your past inertias, whether you're prepared to face a completely different future.

AI isn't the endpoint, but a new starting point. At this starting point, the adaptable survive; the evolved lead.

---

Layout | Yao Nan Image source | Unsplash

Recommended Reading

The flow of money, the rise and fall of people