Today’s theme is printed on the passenger side mirror of almost every car in America. “Objects in mirror are closer than they appear.”

You remember the scene. Jurassic Park. The rear-view shot, the thumping, the water rippling in the cup. And the T-Rex behind them being much closer than the mirror implied.

The mirror is curved on purpose. It shows you more, and in exchange it makes everything look farther away than it is.

That is roughly what has been happening with China and AI.

The obsession in Washington and across our technology companies these past weeks and months has been whether the gap with China on AI is narrowing. And a lot of energy is going into arguing that these objects are not as close as they look. My read is the opposite. They are likely closer than they appear.

Three events today, and all three are the same story from different angles. Moonshot AI and DeepSeek, at the founder level. Unitree Robotics, at the manufacturing level. And AI talent itself, both inside China and inside American labs.

This is the AI Tech Wave seen directly rather than through the glass. Three events, each with my Take, then my Overall Take. Plus a Gadget AI on Apple’s smart glasses privacy problem, and two questions. Let’s get started.

(1) Moonshot AI and DeepSeek. The Founders Up Close.

MP TAKE: There were two interesting profiles out this week. One on Moonshot AI, which shocked the world a few days ago with its Kimi K3 model. Kimi K3 goes right up against the frontier models, Fable 5 and Mythos, Opus 5 from Anthropic, and GPT 5.6 from OpenAI. And it did it as an open-weight model, the largest released to date. Moonshot surprised everyone by landing that close to the leaders, in the open. Its founder is Yang Zhilin, and the FT calls him a rock star founder, which tells you how he is regarded at home.

The other company everyone is fascinated by is DeepSeek, run by Liang Wenfeng, a former hedge fund manager who took his math and quant people, gave them a few billion dollars, and built the first DeepSeek models. The rest, as they say, is history. DeepSeek was in the middle of raising its second round in a month and has now put a pause on it. There was some drama around what was and wasn’t revealed about how they built the operation. Hold that one for a second, because a company that can pause a round of that size is not a company scrambling for validation.

I bring both up because, other than Jack Ma of Alibaba fame, most US observers pay very little attention to the Chinese entrepreneurs who made so much of this possible. Very few here could name Pony Ma, who founded and built Tencent, the other company besides Alibaba that created hundred-billion-dollar-plus global ‘Internet Treasure companies’ over the last twenty five years. Far more, actually, in both cases.

It’s important to note that China’s entrepreneurs and young people have revered and deeply studied the histories of US technology companies, their founding, and their accomplishments ‘changing the world’ globally. Any young person in technology in Chinak or India and elsewhere for that matter, is very familiar with how Bill Gates, Steve Jobs, Jeff Bezos, Larry Page & Sergey Brin, Mark Zuckerberg and others, went on to indelibly change the world.

We haven’t returned the favor. Especially important because their founders, entrepreneurs and young engineers now are doing things worth studying, appreciating and acknowledging,

They are long past the current charges of ‘copying’ and ‘stealing’ our IP. If anything, we’ve been doing it in reverse. Note ‘Reels’ at Meta and ‘Shorts’ at YouTube, which are ‘copies’ of TikTok videos with its unique ‘FYP’ algorithm feeds, all originated by parent Bytedance and its founder Zhang Yiming.

So let me put the names down, because that is half the point of today’s episode.

The ‘Six AI Tigers’ of China. That is what they are called there. The model companies that rose between 2021 and 2024, and the people who built them:

1. Z.ai, formerly Zhipu AI. Founded by Tang Jie and Zhang Peng. Beijing.

2. Moonshot AI. Founded by Yang Zhilin. Beijing.

3. MiniMax. Founded by Yan Junjie. Shanghai.

4. Baichuan AI. Founded by Wang Xiaochuan. Beijing.

5. 01.AI. Founded by Kai-Fu Lee. Beijing.

6. StepFun. Founded by Jiang Daxin. Shanghai.

Note who is on that list. Wang Xiaochuan ran Sogou, one of China’s search engines, before this. Kai-Fu Lee ran Google China, and before that worked at Apple and Microsoft. Jiang Daxin came out of Microsoft Research Asia. These are not kids in a garage. Several of them are second-act founders with decades behind them. Z.ai and MiniMax have both gone toward public listings.

And then, standing outside that grouping entirely, DeepSeek. Out of Hangzhou rather than Beijing or Shanghai, funded from a quant fund rather than venture capital, and the one that rattled the world first. Imagine if Jane Street (epic US quant fund) or DE Shaw (same) in the US spawned a tech company.

Which ironically, DE Shaw kind of did, when Jeff Bezos left a lucrative job and career in the early nineties to move to Seattle with his wife and found Amazon.com.

History rhymes.

Now add the founders beyond the models, because the story does not stop at chatbots. Wang Xingxing at Unitree in robots. Lei Jun at Xiaomi in phones and now EVs. Wang Chuanfu at BYD in EVs. Frank Wang at DJI in drones.

Those are a dozen or so names across the uber Chinese tech/AI companies actually building the physical and digital layers of this wave. If you follow AI closely and can name even half of them without looking, you are in a very small group of Americans. That gap is the distortion this whole episode is about.

Here is the part I most want to land today.

For decades, Chinese technology entrepreneurs have studied our top founders very closely. Not casually. Reverently. Bill Gates. Steve Jobs. How they got there, and then emulating it.

Take Lei Jun, who founded Xiaomi. In 1987, as a computer science student at Wuhan University, he pulled a book off a library shelf called Fire in the Valley: The Making of the Personal Computer. It is the story of Jobs, Wozniak, and the birth of the PC industry in a place he had never been. He has said it changed the direction of his life.

Digest that for a second. 1987. A student in Wuhan reads about Silicon Valley in a library book. He is widely called “China’s Steve Jobs” now, and he wore the black turtlenecks to make the point himself. And the company he built makes not only the Apple-equivalent smartphones and devices in China, but EV cars. If you look at the YouTube videos on the latest Xiaomi sedans, there is no way you would consider another EV if it were available in the United States. Which it isn’t, because we have 100% tariffs on Chinese EVs. You can still buy them in Europe, at relatively steep tariffs.

That arc runs almost forty years. From a book about our founders, to a company competing with ours.

And it has not stopped. It has changed generations. Which brings me to my third point today, and to a house in Palo Alto.

Tariffs and export controls are geopolitical issues, protection of industries, jobs. All real. But the core point I am making is different. Their entrepreneurs are far beyond copying us, or stealing our IP, or distilling our models. They have moved on. They are innovating, and they are creating extraordinary products.

Sources, in narrative order: FT ‘Yang Zhilin, the rock star founder behind China’s Moonshot AI’; Bloomberg ‘DeepSeek said to tell backers of funding pause after viral posts’. For longtime readers: ‘China ramping up in AI vs US open source and beyond’ in AI-RTZ #1153; ‘DeepSeek comes back for Seconds’ in AI-RTZ #1148; and ‘Waiting for AI God-like AGI’ in AI-RTZ #879.

(2) Unitree Robotics. The Manufacturing Up Close.

MP TAKE: Then there are the companies leading entire domains that involve large manufacturing ecosystems. The obvious one is BYD on EVs, built by Wang Chuanfu. The other one, and today’s topic, is Unitree.

Unitree got the cover of TIME, with its founder Wang Xingxing, a former employee at DJI, the company Frank Wang founded. I talked about DJI a few days ago. DJI ended up becoming the Nvidia of drones around the world. Well, Unitree is essentially becoming the Nvidia of robots around the world.

Note the lineage there. The man now leading in humanoid robots trained at the company that took over global drones. That is not a coincidence, it is an ecosystem compounding on itself.

Let me be precise about the category, because this gets muddled very easily. We are talking about humanoid robots. The experimental, still-early class of machine that the TIME piece is about. Not the millions of industrial arms already bolted to factory floors around the world. That is a separate market, far larger and far older, and it is not what is being counted here.

Inside that narrower humanoid category, Unitree is already the volume leader, at more than a quarter of global shipments. And most of those units today go into research and university settings. Not consumer, not industrial. So this is still, for now, largely a research install base.

That does not weaken the point. It sharpens it. Because what Unitree has managed to do, aggressively, is numbers and prices. Compare it to what Tesla is trying to do with Optimus here in the US, with a few dozen robots with barely developed hardware abd software in development. Unitree is shipping in the thousands, and driving the price down hard as it goes.

And that price curve is the story. Cheaper robots mean more of them, in more places, doing more things. It is the Jevons Paradox showing up in hardware rather than in compute. The research install base is the on-ramp to the commercial one, and Unitree is aggressively evolving its manufacturing ecosystem underneath all of it.

That is why they remain the one to watch, respect, acknowledge and use as competitive inspiration. ‘World Cup’ style.

Sources, in narrative order: TIME ‘The Robots Cometh’, by Charlie Campbell, on Unitree, founded and led by Wang Xingxing. For longtime readers: ‘China accelerates in AI Robotics’ in AI-RTZ #858; and ‘US vs China updates on AI Robotics’ in AI-RTZ #873.

(3) AI Talent. In China, and in the United States.

MP TAKE: I am not just pointing at the people at the top, the entrepreneurs who created these things. I am talking about all of the people building these products. And a lot of them have been in the United States.

Which is my third point today. This business of AI talent, not just in China, but here.

The key ingredient for the AI Tech Wave, and I have been saying this for almost four years now across the daily Substacks and the podcasts, is talent. Beyond the chips. Beyond the power. Beyond the memory. Beyond all of the components we constantly talk about on these pages.

Jensen Huang has said China is the second largest market for AI in the world. And China produces the largest share of elite AI researchers of any country on earth. There are a lot of numbers around this, on STEM graduates, engineers, software and hardware, and they are in the show notes if you want the precise ones.

But here is the half of it that gets missed. Most of those researchers now work in the United States. China produces the pool. America employs it. Which means the American AI industry is not competing with that talent so much as it is built on it.

The Chinese are generating STEM and engineering graduates at multiples of what we produce. Not just versus the United States, but versus much of the rest of the world combined. They are terrific, and they work bottom up. They have been inspired by our global technology leadership, and they continue to be, despite our recent moves on immigration and tariffs.

And here is where the arc I started in Take 1 comes back around.

There is a great story in today’s sources about the house Mark Zuckerberg rented around 2004, where he built out the original Facebook. That blue house is now almost a temple for young Chinese engineers, who have rented it for the last three or four years. A constant flood of Chinese founders, twenty-somethings, Gen Z, inspired by what Zuckerberg built.

So think about the two ends of that line. 1987, Lei Jun in a Wuhan library reading about Steve Jobs. 2026, a house full of Chinese twenty-somethings living where Zuckerberg wrote his first code, because being in the room where it happened still means something to them.

Forty years of paying close attention to us. Studying our founders, learning the playbook, revering the people who wrote it.

We have not returned any of that. We could not name most of theirs.

That is the asymmetry I keep coming back to, and it is not really about chips or models at all. We need to start paying attention to their key founders, young and old. Not because it is polite. Because that is how you actually see what is coming.

You cannot measure what you do not see. And you certainly cannot take the measure of talent whose names you never bothered to learn.

Just like Steve Jobs, Bill Gates, Mark Zuckerberg and so many others here.

Sources, in narrative order: The Economist ‘China’s mysterious new billionaires are conquering the world’; FT ‘The vanishing billionaire, how Jack Ma fell foul of Xi Jinping’; Rest of World ‘The Chinese whiz kids of Silicon Valley’; New York Times ‘Why Silicon Valley can’t stop looking over its shoulder at China’. For longtime readers: ‘Meta’s Bingo Card surprises with Manus of China’ in AI-RTZ #952; and ‘US AI Talent hunt includes China’ in AI-RTZ #767.

MP OVERALL TAKE

Chinese AI capability is closer than we think.

And it is not an illusion. It is not an optical illusion. The curve in the glass is not doing the distorting. The assumptions are the core issue.

Notions of competitive races, AI races, space races. They help everybody get their competitive juices flowing, get up in the morning, have a target to do better against. But it is a race worth running only if it makes us better than we would otherwise be. In the end there should also be cooperation and working together.

I laid this out at length over the Fourth of July weekend, in ‘Looking beyond Space & AI Races,’ AI-RTZ #1138. The short version: the race is the headline, and the cooperation is the history. The US and the Soviets ran the moon race, and then shook hands in orbit three years after it ended. Then shared a space station for fifty years after that. That piece is my standing reference on all of this, and today’s episode is the AI version of the same argument.

I keep coming back to the World Cup. The competition was fierce, Argentina against Spain and every match underneath it. And at the end of the day, for the most part, they gave each other hugs and they respected each other. That is what we need with the people we are competing against as a country.

We inspire them, and they inspire us. And we should be watching their competitive spirit far more closely than we do, because they are doing more and more things worth learning from. What they are doing with EV cars. With robots. With drones. With their AI models, open source or closed. These are things we should have been studying closely, and we largely have not, because for decades the US has been the leader.

When I led the internet research effort at Goldman Sachs in the mid-90s, the US was over 80% of the global internet market. Most of the money, the metrics, the growth and the users were here. Today that has flipped. Over 80% of it sits outside the US.

And the growth now is in the Global South. Let me put real numbers around that, because it deserves them. And China top down is more focused on that framework than competing on the ‘East/West’ and ‘Developed vs Developing country’ frameworks.

The ‘Global South’ is roughly 80% of the world’s population. Something on the order of 6.7 billion people, across Africa, Latin America and the Caribbean, most of Asia, and much of the Pacific. Somewhere north of 130 countries.

It is about 40% of world GDP measured in dollars. Measured by purchasing power parity (PPP), which is arguably the more honest measure of what people can actually buy and build with, it is closer to 60%.

And it is roughly two thirds of global growth. Not two thirds of the noise. Two thirds of the actual expansion.

It is also, and this is the part that matters most for AI, young. Africa’s median age is around 19. India’s is around 28. The US is closer to 38. Europe is in the mid-40s. Japan is approaching 50. That is where the next generation of engineers, founders, builders and users is going to come from.

It’s where billions of daily habits are being built. Especially with technologies. Hardware and software. AI and otherwise. And of course, EVs, self driving cars, robots, drones, and all the physical world manifestations of AI technologies ahead of us over the next 25 years.

We have been running while looking in the rear-view mirror at the last 25+ years. All the great things US companies, entrepreneurs and engineers did from the PC to the internet to mobile and now AI.

China is paying attention to every bit of this. Trade, infrastructure, phones, EVs, solar. And now AI models that are cheap enough and open enough to actually deploy in places that will never buy frontier compute at frontier prices.

We largely are not.

And this is where the mirror stops being about China at all.

This was never only a US versus China story. Not East versus West. It is about a far larger part of the world that one side is actively courting and the other side is barely looking at.

Because we are too busy admiring ourselves in the mirror. Checking our own reflection. Our own benchmarks, our own leaderboards, our own valuations. And not seeing the other parties out there. The countries. The companies. The individuals. Not appreciating their talent, their contribution, their hard work, their creativity, their innovation.

And not doing any of it with respect and acknowledgement.

People matter. All of them. Every person, every company, every country wants recognition for what they built. That is where doing more things together than against actually starts.

Which is why this matters particularly at the regulator level. The current US approach to Chinese AI talent, particularly on immigration, and more broadly on how US and Chinese AI counterparts are allowed to interact, is one that needs adjustment. Roughly three quarters of the top China-educated AI researchers already work here. They are not a competing team. They are a large part of our team.

The starting point is simply looking directly at all of this, rather than at the reflections. And not deluding ourselves.

Use the mirrors for the right reflections.

And if you got this far and want to put faces to the names: there is a full list of the companies and founders mentioned today, with links, at the very bottom of this post, under the sources.

GADGET AI

Apple Second-Guesses the Cameras on Its Smart Glasses

MP TAKE: Apple is second-guessing what it wants in its smart glasses. I have talked and written a lot about this.

There is a lot of controversy around Meta’s smart glasses. They sold over seven million of them in 2025 alone, roughly tripling year over year, and they have a whole family of products coming later this year. They have serious competition arriving from Google, in partnership with Samsung and others. Amazon has smart glasses. And Apple is apparently launching in 2027, probably in the fall, with a developer look at WWDC 2027. Bloomberg had a very detailed piece on it.

What Apple is debating is society’s comfort with the cameras in these glasses, the ones taking pictures and video constantly. Given the backlash Meta has seen on privacy, trust and safety, and given regulatory responses in a number of countries saying you cannot wear these in certain venues, schools and public spaces, Apple is trying to figure it out.

There are a lot of ifs and ands in how they might do it. They are experimenting with various forms.

But everyone, all the big tech companies, and companies like Snap, are desperate to avoid a repeat of Google Glass from over a decade ago. Those were very cool gadgets, because the technology allowed it. And the people who started using them were immediately called “Glassholes.” Nobody wants that again. So Apple is being very careful and thinking it through.

Sources: Bloomberg ‘Apple’s Smart Glasses Will Need to Overcome Meta’s Privacy Reputation’. For longtime readers: ‘Apple readies range of AI Smart Glasses’ in AI-RTZ #1055.

Q1. What is my latest stand on using the camera features of the Meta AI glasses?

Answer: I use them sporadically. Honestly, not as much as I probably should.

And they are awesome, because of the point of view you get. Whether you are running with them or just walking around, you get a centered, first-person view rather than the kind of view you would get holding up a camera or a smartphone. That is the most attractive piece of the whole thing.

Q2. What is my biggest issue with using the camera on the glasses?

Answer: It is the comfort level of using it around people you don’t know.

It is fine to use with friends and family, with their permission, taking better pictures of the kids and the parties and so on. It is another thing entirely to take pictures of people out in the open without their permission. And that is exactly where I tend not to use these things at all.

It remains a very important dynamic, and who knows how it shakes out for the big companies and the smaller ones rolling these things out over the next twelve to eighteen months. To be determined. But it is an important issue, especially in light of this new set of experiments at Apple.

WRAP

Today’s AI-RTZ #1160 Markets Looking Beyond Nvidia. It is on the current relative underperformance of Nvidia versus its chip peers, and why the market may be under-appreciating what is still ahead for it.

The short version. Nvidia is up about 11% on the year while AMD is up over 140% and Micron over 200%. The Information argues the stock is priced as though everything that can go wrong will. I lay out six things I think the bear case is missing, drawn from what I have written here over the past year. The custom silicon everyone is building is a complement, not a replacement. Jevons Paradox is enlarging Nvidia’s market, not shrinking it. Self-driving is the same under-appreciation, already visible in one vertical. The market pays trillion dollar premiums for narratives and nothing for the platform already shipping. Jensen is the picks-and-shovels supplier to every side. And Nvidia has turned technical leadership into open source standards leadership.

Which ties directly to today’s episode, since a good deal of what Nvidia is arguing for in Washington is the same open posture toward China that we have been talking about here.

AI Ramblings Daily on AI-RTZ is here to think through AI and reset. Together.

Tomorrow, ARD 128 and AI-RTZ #1161.

Thanks for joining us today, AI Curious Folk. Stay tuned.

— MP

Full Source Reading

For the broader context, see the canonical sources for ARD 127, in today’s narrative order:

Event 1. Moonshot AI and DeepSeek, the Founders Up Close

• FT, Yang Zhilin, the rock star founder behind China’s Moonshot AI

• Bloomberg, DeepSeek said to tell backers of funding pause after viral posts

• Wikipedia, the Six AI Tigers

• Britannica, Lei Jun

• Quartz, the book about Steve Jobs that changed Lei Jun’s life

Event 2. Unitree Robotics, the Manufacturing Up Close

• TIME, The Robots Cometh

Event 3. AI Talent, in China and in the United States

• The Economist, China’s mysterious new billionaires are conquering the world

• FT, The vanishing billionaire, how Jack Ma fell foul of Xi Jinping

• Rest of World, The Chinese whiz kids of Silicon Valley

• New York Times, Why Silicon Valley can’t stop looking over its shoulder at China

Gadget AI. Apple’s Smart Glasses Camera Question

• Bloomberg, Apple’s Smart Glasses Will Need to Overcome Meta’s Privacy Reputation

Clips from today

Clip 1. Meta Glasses: The Privacy Problem

Watch on YouTube Shorts

What’s my biggest issue with the camera on the Meta glasses? It’s the people who never agreed to be in the shot.

MP Take: I use them sporadically, honestly not as much as I probably should. And they’re awesome because of the point of view you get. Whether you’re running with them or just walking around, you get a centered, first-person view rather than the kind of view you’d get holding up a camera or a smartphone. The negative, and my biggest issue, is the comfort level of using it around people you don’t know. It’s fine with friends and family, with their permission. It’s another thing entirely to take pictures of people out in the open without their permission. And that is exactly where I tend not to use these things at all.

Clip 2. China’s AI Talent Surge

Watch on YouTube Shorts

The core ingredient in all of this isn’t the chips, the power, or the memory. It’s the talent.

MP Take: For almost four years now, across the daily Substacks and the podcasts, I’ve kept coming back to the same point. Beyond the chips, beyond the power, beyond the memory, the key ingredient for the AI Tech Wave is world-class AI talent. Jensen Huang has said China is the second largest AI market in the world. And China produces the largest share of elite AI researchers of any country on earth. China is generating STEM and engineering graduates at multiples of what we produce, not just versus the United States, but versus much of the rest of the world combined.

Clip 3. Chinese AI: Closer Than It Appears

Watch on YouTube Shorts

Objects in mirror are closer than they appear. So is Chinese AI. And it is not an optical illusion.

MP Take: My overall take on all of this is that Chinese AI capability is closer than we think. It’s not an illusion, and it’s not an optical illusion. It’s a race only if the race helps us be better than we would otherwise be. The mirror analogy only goes so far, so keep in mind how close or far they actually are, because a race is a race. But the core thing is that we inspire them and they inspire us. And we should be more inspired by their competitive spirit than we currently are, because they are doing more and more things worth being inspired about. What they’re doing with their EV cars. Their robots. Their drones. Their AI models, open source or closed.

About AI Ramblings Daily (ARD), and AI-RTZ

Both are daily. Both are free. Both are about AI. But they’re different mediums carrying different messages.

AI-RTZ is the morning text, a deeper written take on one idea, published by at least 5 AM EST. Today: post #1160.

AI Ramblings Daily is the afternoon video + podcast, my ad hoc takes and perspective on the day’s AI issues and news flow, around 20 minutes, with short 1-2 minute clips for quick topic views. Today: episode #127.

Subscribe to either or both on michaelparekh.substack.com. They run as separate Sections you can opt into or out of.

Links used in today’s show (already embedded inline above; listed here for reference)

Take 1. Moonshot AI and DeepSeek, the Founders Up Close:

• FT, Yang Zhilin, the rock star founder behind China’s Moonshot AI

• Bloomberg, DeepSeek said to tell backers of funding pause after viral posts

• AI-RTZ #1153, China ramping up in AI vs US open source and beyond

• AI-RTZ #1148, DeepSeek comes back for Seconds

• AI-RTZ #879, Waiting for AI ‘God-like’ AGI

• Wikipedia, the Six AI Tigers

• Britannica, Lei Jun

• Quartz, the book about Steve Jobs that changed Lei Jun’s life

Take 2. Unitree Robotics, the Manufacturing Up Close:

• TIME, The Robots Cometh

• AI-RTZ #858, China accelerates in AI Robotics

• AI-RTZ #873, US vs China updates on AI Robotics

Take 3. AI Talent, in China and in the United States:

• The Economist, China’s mysterious new billionaires are conquering the world

• FT, The vanishing billionaire, how Jack Ma fell foul of Xi Jinping

• Rest of World, The Chinese whiz kids of Silicon Valley

• New York Times, Why Silicon Valley can’t stop looking over its shoulder at China

• AI-RTZ #952, Meta’s Bingo Card surprises with Manus of China

• AI-RTZ #767, US AI Talent hunt includes China

Gadget AI. Apple’s Smart Glasses Camera Question:

• Bloomberg, Apple’s Smart Glasses Will Need to Overcome Meta’s Privacy Reputation

• AI-RTZ #1055, Apple readies range of AI Smart Glasses

Companion text:

• AI-RTZ #1160, Markets Looking Beyond Nvidia

Standing reference on US/China races and cooperation:

• AI-RTZ #1138, Looking beyond Space & AI Races. From Cold Wars 50+ years ago to US/China AI/Tech ‘Wars’ today

The companies and founders in today’s show

A reference list, since most of these names are not household ones in the US, and that is rather the point. Two good roundups first, then the individuals.

The Wire China, Who’s Who in China’s AI Industry

The Six AI Tigers

The Six AI Tigers, the model companies:

• Z.ai, formerly Zhipu AI. Founded by Tang Jie and Zhang Peng. Beijing.

• Moonshot AI. Founded by Yang Zhilin. Beijing.

• MiniMax. Founded by Yan Junjie. Shanghai.

• Baichuan AI. Founded by Wang Xiaochuan, formerly of Sogou. Beijing.

• 01.AI. Founded by Kai-Fu Lee, formerly of Google China. Beijing.

• StepFun. Founded by Jiang Daxin, formerly of Microsoft Research Asia. Shanghai.

And beyond the models:

• DeepSeek. Founded by Liang Wenfeng. Hangzhou, out of a quant fund.

• Unitree Robotics. Founded by Wang Xingxing. Humanoid robots.

• Xiaomi. Founded by Lei Jun. Phones, and now EVs. Here is the book about Steve Jobs he read in 1987.

• BYD. Founded by Wang Chuanfu. EVs.

• DJI. Founded by Frank Wang. Drones.

• ByteDance. Founded by Zhang Yiming. TikTok, and the feed algorithm everyone copied.

• Tencent, founded by Pony Ma. Alibaba, founded by Jack Ma. The generation before.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here.)

Subscribe to AI: Reset to Zero for daily AI Ramblings + Sunday Bigger Picture posts.