The Rise of China’s Electric Vehicle Industry

The Machine Behind BYD

How did China grow its EV industry? Hint: it wasn’t really central planning from Beijing. It was a magical concoction of entrepreneurs and local officials using their guanxi to dance around giant SOEs and central-government regulation, leverage private and foreign capital, and just be really excellent business people. That’s led to our situation in 2026 — world-beating companies that have a ton of overcapacity and are set to take over the planet.

To discuss, we have ANU Professor Fengming Lu, who recently wrote a fantastic paper in The China Journal, titled The Rise of China’s Electric Vehicle Industry: Strategic Alliances Between Local Governments and Private Capital, alongside Irene Zhang, contributor and analyst at China Talk. We also get into the lessons this story holds for other strategic industries like robotics and AI — plus a little detour into the Chinese podcasting ecosystem.

Our conversation covers:

• China’s automotive transformation from 130 small carmakers in the 1990s to the world’s most competitive EV industry.

• The rise of Chery, Geely, and other Chinese auto champions — how local connections, foreign technology, and creative financing helped them break into an industry dominated by state-owned enterprises.

• The downside of success — how China’s EV boom created an overcapacity problem.

• The role of guanxi and elite politics in deciding who gets to build cars.

• When Chinese EVs go global — record exports, overseas factories, trade tensions, and whether the same local-government playbook will shape the next wave of industries, from robotics to AI.

Listen now on your favorite podcast app.

Local Beginnings of the Chinese Auto Industry

Irene Zhang: I had an amazing time reading your paper, and I found it incredibly interesting to learn the long history of how the global wave of Chinese EVs we’re seeing today actually came about. I’d like to start by hearing the story from your side. How did you and your co-author come up with the idea for this paper — how did you get interested in EVs as a research topic and decide to write it?

Fengming Lu: It’s connected to my own research agenda. I started doing research on EVs a few years ago, roughly during the final years of COVID. After that, I was mostly doing fieldwork, and I began writing papers and working on book projects last year.

The original idea was to write some of the first papers for my EV research project. My co-author and I knew each other very well, and we’d come across several questions from senior scholars who study Chinese government and the Chinese economy. One of the first puzzles they raised was this: over the last ten or twenty years, especially after 2012, you see the rise of SOEs — state-owned enterprises — in many corners of the Chinese economy. But the automotive sector, or the EV sector, is an exception. It’s probably not the only one. If you look at the tech sectors in China, most are dominated by private firms rather than SOEs. That’s a pretty interesting puzzle for anyone interested in China.

My co-author and I thought it would be a good idea to start with a paper about this. He added some other cases from his own expertise, like pharmaceuticals and steel manufacturing, which also fit the pattern: the private sector, building on collaboration with Chinese local governments, excels at the expense of traditional, more established SOEs.

Jordan Schneider: This was the most fascinating arc in your paper. On one hand we have Xi Jinping, the biggest SOE cheerleader we’ve had in a few premiers. Yet perhaps the most globally impressive, successful industrial leapfrogging has taken place in an industry where the SOEs were basically stuck, doing nothing.

And then two different development pathways were tried — partnering with foreign automakers, and building it themselves — and they were fine, but nothing all that impressive. Then all of a sudden, as you explore in this paper, you get this really interesting mix of private capital, entrepreneurs, local governments, and regional incentives, all dancing around whatever space the central government and the big SOE carmakers left open, creating the incredible flourishing we’ve seen over the past ten years.

Let’s do a bit of a timeline of the Chinese car industry. Why don’t you take us back to the ’90s? Where can we start to see the seeds of the EV boom in that period?

Fengming Lu: Absolutely. And thank you, Jordan and Irene, because you really read my paper rather than just the abstract.

Back in the ’90s, the Chinese automotive industry was in a very different shape from what it is now. If you’re talking about passenger vehicles — cars, SUVs, and so on — the market was only about half a million sold every year. It was much smaller, comparable to Poland back then, or Vietnam today.

For the Chinese government, the main concern was a legacy of the Maoist era and the economic decentralization of the 1980s: the automotive industry was super fragmented. As far as I remember, every province had its own carmaker except Tibet. All twenty-nine or thirty others had their own. By the mid-1990s, someone did the count and China had about 130 carmakers across the country — roughly the same number of carmakers you’d find at the peak of the EV boom, even though back then their combined output was tiny by comparison.

Jordan Schneider: What I thought was fascinating, Fengming, is that it actually made sense to have a local manufacturer in whatever corner of your province. You had these Maoist controls on who got which automobile, so if you wanted any cars in your region, you kind of had to make them yourself. On its face it seems absurd to have a company that makes 500 cars a year, but it may actually be economically productive if the difference is between having 500 cars or trucks versus having none at all.

An ad for the Yemingzhu YMZ5010X (made in Chengdu), from the January 1992 issue of the Sichuan Automotive Journal. Source.

A 1992 brochure showing the YJ2020 made by Yanjing Auto (燕京汽车), based in Beijing. Source.

Fengming Lu: To some extent, it’s similar to the British automotive sector, with all those countless brands and carmakers. What worried the Chinese government was a shock that began in the late 1970s and early 1980s, when they started negotiating with European carmakers, especially Volkswagen. Volkswagen — and many others — basically told them that to make a carmaker efficient, you need to manufacture at least 150,000 cars a year. That’s the minimum economy of scale.

The Chinese government, the regulators, and the policymakers kept this magic number in mind: 150,000 cars a year. So they did some simple math, especially in the 1980s. The market was below one million, which meant it could only support four to six large-scale carmakers. Very simple math.

That’s why, starting in the late 1980s, the Chinese government came up with a concept that’s very famous in the automotive sector: the “three bigs and three smalls” (三大三小). Back in the 1990s, those were the legitimate carmakers you could see on Chinese roads — though of course not the only ones. FAW (一汽) had a joint venture with Volkswagen. Dongfeng (东风), still known as the Second Automotive Works, had one with Citroën. Shanghai already had its joint venture with Volkswagen as well. There were smaller ones too, like Beijing Jeep (北京吉普) and Guangzhou Peugeot (广州标致). And Tianjin Xiali (天津夏利), which had a license from Daihatsu to make smaller Japanese cars.

That wasn’t the only list, either. Later, some carmakers with defense or arms-manufacturing backgrounds — all state-owned, of course — somehow sneaked into the game too, like Chang’an . But it was still a very restrictive list. And elite politics played a role: many of those carmakers pulled various strings to get onto the list.

What you see in the 1990s is a set of approved carmakers in China, but plenty of other provinces and cities still wanted to get into the game. Back then, China hadn’t joined the WTO, so tariffs were very high. In the early 1990s, tariffs on cars ran around 150 to 200 percent, and later came down to somewhere between 80 and 120 percent. So for many local governments and firms, making cars was extremely profitable. An easier route, of course, was to smuggle cars — especially for places like Guangdong, Fujian, and, very interestingly, Shandong, Liaoning, and Jilin, because they were close to South Korea.

But some other cities tried to figure out their own ways into automotive manufacturing. Several of the giant carmakers in China today trace their roots to the 1990s, when they started from very humble beginnings and were later either created by local governments or collaborated with them — companies like Geely (吉利), which today owns Volvo Cars and runs its own brands like Geely and Zeekr (极氪). And Chery (奇瑞). Chery is a very interesting company. They’re rising so quickly now in the European, British, and Australian markets, yet people outside China know so little about them — even people outside Anhui Province know so little.

Jordan Schneider: Great, let’s start with Chery, then. Why don’t you tell the origin story of how it grew starting in the ’90s?

Fengming Lu: It’s such an interesting company. Originally, it was a state-owned enterprise — and it stayed one through the ’90s and most of the 2000s — but it wasn’t owned by any large city, large province, or the central government. It was the state-owned enterprise of a city called Wuhu (芜湖). Very few people have heard of Wuhu outside China, and even inside China not many have. It sits in the middle of Anhui Province, about 500 to 600 kilometers (~400 miles) from Shanghai.

In the early 1990s, the city and the province weren’t in great shape. Anhui had been lagging behind its neighbors — Jiangsu, Zhejiang, and of course Shanghai — since the start of the reform era, in many ways. So they’d been looking for new industries to boost their economy.

I went there for in-depth fieldwork, and combined with other materials I collected, I can reconstruct the story. In the early 1990s, Wuhu was searching for a new industry. They tried many things, and then they noticed a township-and-village enterprise in a village near the city. If I remember correctly, it had started collaborating with an arms maker in Henan Province — they’d somehow made a connection over there, found some engineers, and started building cars based on kit parts from South Korea. They were manufacturing on a very small scale: about 3,000, 4,000, 5,000 cars a year.

Jordan Schneider: Before we go on, can you do two sentences on TVEs? If people have heard of them, they mostly think of them as the thing that got agriculture going in the ’70s. How do township-and-village enterprises end up turning into small carmakers and manufacturers?

Fengming Lu: In the 1980s you saw a lot of small township-and-village enterprises flourishing across China. But starting in the late ’80s and early ’90s, some of them were hit very hard, because their products weren’t technologically advanced or sophisticated. There was also a harsh crackdown on TVEs by the central government after 1989. So a lot of those enterprises were looking for new directions.

This particular TVE found a new way: making cars. It was immensely successful. Their cars weren’t of very good quality, as far as I can tell, but they sold for around US$10,000 — roughly 70,000 or 80,000 Chinese yuan — much cheaper than what the “three bigs and three smalls” state-owned carmakers offered. They made quite a lot of money and became the first company in Wuhu to exceed 100 million yuan in revenue.

So Wuhu’s officials started asking why this company made so much money, and the answer turned out to be cars. The leadership of Wuhu discussed it with the leadership of Anhui, and they decided to give car-making a try. One provincial official put it roughly this way: maybe the party and the government can do something in car-making.

They first pulled strings with one of China’s largest state-owned carmakers, FAW — First Automotive Works. FAW’s general manager at the time was a man named Geng Zhaojie (耿昭杰), still a very respected figure in the Chinese automotive industry for many things he did. He happened to be from the area near Wuhu. The leader of Wuhu befriended him through the National People’s Congress, China’s rubber-stamp legislature. It may not be very meaningful as a legislature, but it’s quite meaningful as a social club. They also befriended another entrepreneur nearby.

FAW in 1956. Source.

So they thought about establishing a carmaker in Wuhu, but it didn’t end well. Government officials in Wuhu told me FAW was so arrogant toward them because FAW saw itself as the grand “eldest son” of the Chinese automotive industry. In administrative rank, FAW was more senior than the city of Wuhu, and it had been the cradle of many of China’s political leaders. So FAW felt it didn’t need to listen to Wuhu. All FAW ended up doing was making peripheral components — chassis for trucks and so on — in Wuhu.

The most telling story officials told me was that the managers FAW sent would smash into the conference room in the middle of Wuhu’s own meetings, shouting, “Is your party secretary here? I want to talk to him.” It was very rude, and eventually Wuhu couldn’t stand it.

At that point, they had a windfall. They sold a cement works to a nearby company and somehow got 200 million Chinese yuan — a large amount back then. That cement works later became Hailuo Cement (海螺水泥); if you follow the Chinese stock market, it was once one of the largest cement companies in China and a Fortune 500 company. They used that money — plus bank loans arranged through the provincial government — when they heard that some engine-manufacturing assembly lines were on offer in Britain. So they organized a trip to Wales, where Ford of Britain was selling an engine-manufacturing line. What they’d learned from FAW was that without your own engine, you can’t make your own cars. Later, they recruited engineers from FAW and elsewhere, and started their own carmaker: Chery.

How they picked their first product is also interesting. It was reverse-engineered from the SEAT Toledo — essentially the Spanish version of the Volkswagen Jetta. The Jetta was also sold in the US in the ’90s, but in China it was one of the very few affordable family cars. Chery decided to go head-on with the Jetta, partly because they had engineers from FAW who knew how to build that kind of car.

Later, in the 2000s and 2010s, Chery gradually became more of a private firm — through a management buyout, and more recently an IPO in Hong Kong. The governments of Anhui and Wuhu still hold about 30 percent of the stock, but they’re no longer the majority shareholder.

That’s the story of Chery, and it’s a great example of how local governments were essentially blocked from making cars. Wuhu had no central-government license to build cars, and for the first few years, they could only sell within Anhui Province. They knew the Anhui market alone could never sustain a carmaker, so they tried several workarounds. One thing I remember — I was a kid then, maybe ten years old, and a car fan — was that they sold cars in Chengdu, an emerging market for car buyers, with Anhui license plates. If you wanted one of the first batch of Chery cars in Chengdu, you had to register it with an Anhui plate. The first people to complain were the Chengdu police, who couldn’t stand it and told buyers they had to register their cars locally.

Manufacturing from Scratch

Irene Zhang: If you were Wuhu, you had a TVE that manufactured vehicles. But for some of the other cities you discuss in the paper, I’m curious how they even got the technology to make cars in the first place. How do you make a car if you’ve never made one, as a city trying to build this kind of relationship?

Fengming Lu: In the ’90s it was actually very hard, because for many of them, all they could start with was reverse engineering — basically copying existing models. Some could afford to buy a bit of technology from carmakers: licenses to build engines, or some blueprints. But few could start from scratch. Even the Chinese joint ventures were still learning how to manufacture cars properly throughout the 1990s. So most of the firms that began as these strategic alliances between local governments and private capital were building relatively low-quality cars, much like the humble beginnings of Japanese and Korean carmakers.

From the 2000s onward, things changed quite a bit. One of the most interesting changes was recruitment. Many firms had started by recruiting talent from within Chinese carmakers and joint ventures — excellent manufacturing engineers who didn’t necessarily know how to develop or engineer cars from scratch. So they began recruiting from overseas: Detroit, Japan, and Europe. Many of those recruits, of course, were ethnically Chinese.

Another interesting change was the capital market. I’m working on a book project, trying to explain why the Chinese EV and automotive sectors have surprised so many people. Ten years ago, they were seen as a joke by many observers, but now they’re one of China’s key pillar industries. I’d say there are three pillars. One is the central government’s industrial policy, which we’ve already discussed a lot. The second is local government, which is central to my paper. And the third is the capital market.

Let me tell you something interesting. Can anyone guess which company was the first from China — even from Greater China, including Hong Kong and Taiwan — to IPO on the New York Stock Exchange or NASDAQ?

Jordan Schneider: Oh man, that’s tough. Can you give us a year?

Fengming Lu: It was 1992. Even before TSMC.

Irene Was it an energy company?

Fengming Lu: Not an energy company — the state-owned enterprises went to overseas markets much later. It was actually a car company. More precisely, a minivan maker, called Brilliance Auto (华晨汽车).

A lot of things happened to that company later, but they were the first to go overseas to sell their stock. They raised somewhere between fifty and sixty — actually about seventy-something — million US dollars on the New York Stock Exchange. Their tagline was that they were the first company seeking an IPO from a communist state, which was sexy enough in a lot of ways. They raised around US$70 million, five years before TSMC. That was huge.

For most other Chinese carmakers back then, all of them essentially state-owned, the only way to raise tens of millions of dollars — or even US$100 million — was bank loans from the state-owned banking system, or in some cases from Western banks on very restrictive terms. The Chinese carmakers found this new way. And later, of course, the Chinese internet companies — Sina, Sohu, and then Alibaba and others — really maximized this model.

Jordan Schneider: That’s basically the story of the ’90s. If you wanted to do something more creative than the SOEs, you needed some other type of guanxi to get around the regulatory restrictions and local selling barriers. Maybe you got drunk with someone at an NPC session and they hooked you up; maybe a mayor happened to be really into cars. It was incredibly difficult, because you had the centralized banking system as well as the large SOE carmakers, who had no interest in your small town and its big dreams.

But coming into the 2000s and 2010s, you start to have this other pathway — raising foreign capital, or pooling money from lots of local banks. We can turn now to Geely as the exemplar of just how creative you can be, going around official Beijing priorities to build what’s now a world-beating firm.

Fengming Lu: Absolutely. Geely has been moving aggressively in the West recently. I think both of you are based in San Francisco, right? You can see some of the new Waymo vehicles — the van-type ones. Those were made by Geely.

The Geely Ojai. Source.

It’s a fascinating story. We just talked about Brilliance Auto, which seems totally unrelated, but here’s the connection: around 2003 or 2004, Geely recruited Brilliance Auto’s CFO to be its own CFO. He brought all the know-how on managing the stock market and raising capital into Geely. From that point on, Geely made a lot of smart moves. They were among the first firms to IPO in Hong Kong.

The highlight for Geely was leapfrogging a huge barrier in automotive manufacturing, engineering, and technology by acquiring Volvo Cars in 2010 — Volvo Cars specifically, not the Volvo Group — from Ford. It’s a great illustration of how Geely combined the capital market and local governments to do something big.

I interviewed some of the people who participated in that deal; it stretched from 2007 to 2010, taking Geely almost three years. One of my favorite details — which I also posted on Twitter — involves the lead investment banker, a woman named Jennifer Yu from Rothschild & Co., the European investment bank. Around 2008, she was pitching her colleagues on committing to help Geely. She told them that although Geely was a mouse trying to swallow an elephant — the elephant being Volvo Cars — it was a deal worth doing. At the time, Geely’s revenue was only about 10 billion Chinese yuan, yet acquiring Volvo would cost over US$1 billion. They simply didn’t have that much cash in the bank; they had to find a new way.

Jennifer’s pitch was that, although it was a mouse trying to swallow an elephant, it was backed by a gigantic dragon. She was probably referring to China’s foreign reserves. But that’s not how things work in China — you can’t just go to Beijing and say, “I’m a company trying to acquire an established foreign carmaker. Can you give me a billion or two from your foreign-reserve account?” You need the backing of the banking system, a kind of endorsement, to pull off the deal.

So Geely’s idea was to raise some money from the Hong Kong stock market — but that wasn’t enough — and then lobby for support from various Chinese municipal governments. Starting in 2008: Ford was initially unwilling to sell Volvo Cars, but after the financial crisis they said that if Geely could take good care of Volvo and offer a reasonable price, they could do the deal, on the condition that Geely raise the money itself. So Geely’s most important job throughout 2009 was getting roughly US$2 billion together.

One of the investment bankers on the deal told me Geely talked with as many as 60 cities in China. China has about 300 prefecture-level cities, so that’s a fifth of them. What Geely showed municipal officials was the next-generation Volvo lineup — the XC60s, XC90s, and so on that you see today. The pitch was: “I can let you manufacture some of these models, but in exchange, five fingers.” Five fingers meant 500 million Chinese yuan — about US$70 or 80 million back then. That was the ticket to get on board. With credit from your local banks, or your investment as equity into the acquisition vehicle, “if we assemble four or five cities into this game, Geely can make the system work.”

Initially, the biggest supporter — something I didn’t include in the paper, but probably will in a future book — was Beijing. Surprisingly, Beijing wanted Volvo. Geely even showed Ford’s CEO a letter from one of Beijing’s municipal leaders to demonstrate that it had Beijing’s backing. But in late 2009, Beijing withdrew, because it had its own SOE, BAIC (北汽). BAIC told Beijing’s leaders not to back a competitor within the city — to support BAIC as the sole municipal champion and not stir up trouble by finding a rival.

So Geely — and especially its founder and CEO, Li Shufu (李书福) — was furious and frantic. How could he raise all that money before the deadline? Ford had set it at, if I remember correctly, December 20th — the last business day before Christmas. Could he get the money from Chinese cities before then? Long story short, Geely lobbied many cities and eventually found support from several that wanted an automotive industry. Chengdu, for instance, had been struggling to land an anchor carmaker, so it was very pleased to see Geely and put in 500 million yuan as a bank loan. Some other cities injected funds as equity into a special-purpose vehicle for the acquisition — very interestingly, Daqing, famous as an oil city. Daqing had a lot of cash from drilling but knew it would run out of oil soon, so it wanted a Volvo factory. Shanghai was involved as well.

Jordan Schneider: Let’s recap. In the ’90s, if you were Shanghai you got a joint venture because Jiang Zemin and Zhu Rongji liked and trusted Shanghai. If you were Wuhu, you had this weird connection — Jiang Zemin was sort of from there too — so they let it go. But by the late 2000s and early 2010s, you could be more entrepreneurial.

I’m curious where the central government’s head was on all this. Who knows how much money was changing hands behind the scenes to get things approved — but how happy were the central planners with this entrepreneurial activity, or were they still annoyed?

Fengming Lu: Several things were going on. With Geely’s acquisition of Volvo, the central government’s initial instinct was still to favor the SOEs. What made the deal work was that Geely was very crafty about it: it didn’t move until it had unwritten approval from the NDRC, the National Development and Reform Commission.

If you look back at the news from late 2007 and 2008, the earliest candidates to buy Volvo were companies like Chang’an, a major SOE that also had a joint venture with Ford — a natural partner. But all those state-owned enterprises eventually said no, claiming they didn’t have enough money or that buying Volvo wasn’t a priority. Once the NDRC made sure none of the SOEs were buying Volvo, they told Geely it could go.

There were other Chinese private automakers trying similar games. One of the craziest — a company from Sichuan — tried to buy Hummer, the maker of the US Army’s Humvee. They didn’t get approval, but they announced the deal anyway. The regulators in Beijing reportedly learned about it from the newspaper and TV — “We didn’t hear about this” — and killed it. So there’s nuance here.

The main point is that there was still a bias toward SOEs among Chinese regulators back then. But toward the end they were also happy to see a private carmaker acquire a foreign company if it could boost China’s technology. The catch was, “We’re not providing any financial support. You need to sort that out on your own.” For Geely, that meant dealing with all those local governments first. Even after the NDRC’s approval, the policy banks — like China’s Export-Import Bank — said no. It wasn’t until Geely had secured all the funding that China Development Bank decided it could offer some loans for later construction and plans. They needed to see that Geely had secured everything first.

The Birth of the Chinese EV Ecosystem

Jordan Schneider: So we’re up to the mid-2010s, when electric vehicles really start being part of this story. There’s a big debate — in the literature and among policymakers — about who gets credit. Is it the entrepreneurs, the local governments, or the central government? Why don’t you narrate the birth of the Chinese EV ecosystem and tell us who you think deserves the plaudits for it being born the way it was?

Fengming Lu: My paper tells one aspect of the story — an important part — but I’m writing a book to tell the fuller version. All of them deserve some credit.

We’ve already heard a lot of praise for the central government’s industrial policy: that China had a master plan to build EVs starting in the early 2000s. Parts of that are true. China’s support for EVs has been remarkably consistent over the last 20 years, which makes a real difference compared to all the back-and-forth in the US recently. But it’s not the only part of the story. A decade ago, many people weren’t optimistic about the sector, even though the state was already giving big subsidies. What’s been missing from much of today’s discussion is the role of local governments — and not just private entrepreneurs, but the capital market too.

The most important case is BYD. BYD started as a humble battery maker. Initially, they weren’t even making batteries for cell phones — they made batteries for toys, electric tools, and emergency lights in the 1990s, and later for cell phones. They entered automotive manufacturing in the early 2000s, largely because they’d raised a lot of money on the Hong Kong stock market and figured car-making was lucrative. One big difference was that they were committed to EVs because they could make batteries.

The early years were rough. BYD started making plug-in hybrids around 2008, but the major breakthrough came in the early 2010s. It was made possible in part by central-government industrial policy, but that wasn’t the only factor.

Here’s the story. A lot of discussion of the Chinese EV sector focuses on the early pilot projects. The idea was that EVs would be too expensive for private buyers when batteries were so costly, so governments should buy them first, as buses, taxis, or cleaning vehicles. Public-use fleets, even in small numbers like a few thousand a year, could create a market from scratch.

That’s why China ran two demonstration programs — during the Beijing Olympics and the Shanghai Expo — to showcase its willingness to adopt EVs and other clean vehicles. Around 2011, China also launched the “Ten Cities, Thousand Vehicles” project, initially among ten cities and later among thirty.

The name tells the story. The Chinese government said that for cities willing to adopt more EVs, the goal was to have at least 1,000 electric vehicles running as buses, taxis, and so on within two or three years. It’s a very Chinese way of designing policy: the central government dangles bait for local governments, then assesses performance after two or three years.

BYD realized that many local governments wanted into these projects — eventually 30 cities joined the pilot, far more than ten. The selected cities wanted to maximize their EV usage as a kind of political performance. So BYD chose early local-government partners carefully. The most interesting cases are its hometown, Shenzhen, and the city of Changsha. Changsha had just sold a coachbuilder to BYD around 2008–2009, and BYD worked closely with both cities on the Ten Cities, Thousand Vehicles project.

If you visited Shenzhen around 2010 or 2012, you’d find that overnight, all the taxis had been replaced by BYD EVs. To be frank, they weren’t of great quality — many people told me so — but the Shenzhen government insisted on using them. Many other cities copied this once BYD agreed to invest or build factories there: Changsha, Taiyuan around 2015 or 2016, and others. Overnight, all the taxis became BYD EVs.

BYD also created strong incentives for local governments to succeed in these programs. Ilaria Mazzocco of CSIS has discussed some of this, but there are more fascinating stories. BYD set up its own financing companies to help taxi drivers and bus companies buy EVs. And although Shenzhen didn’t host the Olympics or the Expo, it was holding a much smaller event — the Universiade, essentially the Olympics for university students. The Shenzhen government took it very seriously and decided to copy the Beijing and Shanghai model with its own EV demonstration program. They asked BYD to deliver EVs, and Shenzhen eventually put 2,011 EVs into service for the event — BYD contributed a quarter of them.

It was a humble start, and BYD copied the model in many other cities. Throughout the 2010s they invested in city after city. Today, they have major assembly plants and other facilities across more than a dozen Chinese cities. Their strategy was interesting: they tended to choose second- and third-tier cities, especially in central or western China, where labor was much cheaper.

They also had this exchange to offer: “We’ll replace all your taxis and buses with our EVs.” To smooth collaboration further, BYD has a dedicated rail-transport division. They’d tell local governments, “If you want more from us, we can build a light-rail project for you.” It’s a mass-transit rail system the local boss could point to as a big achievement.

That was one secret behind BYD’s quick expansion after the late 2010s. Actually, “succeeded” is too strong for that period — they sustained themselves through the 2010s. Back then, BYD was much smaller, with the nickname “B-thirty-thousand” (B三万 B sānwàn), because they could only sell about 30,000 vehicles a month. Now they sell around 300,000 a month — about ten times as many. They survived those ten years largely by selling EV buses and taxis to cities, in China and abroad. The key link was that by helping local governments hit their policy targets, BYD simultaneously expanded its manufacturing and market share.

BYD’s first experimental electric vehicle in 2002. Source.

Irene Zhang: This conversation just unlocked a deeply buried memory. I lived in Shenzhen for a few years as a kid in the 2000s, and I remember going back around 2017 after a long absence. I got out of the train station with my family, walked toward the taxi lane, and every single cab was the exact same make. I didn’t know what it was back then — I was just a kid thinking, “Wow, they exchanged all the cabs.”

Fengming Lu: Yes — there’s a very good reason for that.

The Overcapacity Problem

Irene Zhang: This is a good segue toward the present: this entire machine of an industry going abroad and reshaping the global automotive market. Do you want to talk about the capacity-expansion side and how Chinese carmakers ended up shaping the industry writ large?

Jordan Schneider: And also, to what extent is the overcapacity something the central government is happy or upset about?

Fengming Lu: Overcapacity has been chronic in the Chinese economy, especially in manufacturing. One key driver is this combination of private capital and local governments. It happened before in steel and other sectors — and now in autos.

What really happened in the 2010s is widely seen as a turning point. In 2014 and 2015 — the peak of “mass entrepreneurship and innovation” (大众创业、万众创新) — the Chinese government lifted the restriction on private capital, and even private startups, building cars. That’s why China has not only the companies you see today, like NIO (蔚来), XPeng (小鹏), and Li Auto (理想) — what people call the “new forces” (新势力) — but also a wave of now-bankrupt firms. I did the count for another paper I’m writing: there were 60 or 70 EV startups in China after 2015. That’s crazy. In the US you have Tesla, Rivian, Lucid, maybe Fisker — five or six. In China, 60.

Many were purely private and weren’t very successful because they ran out of cash early. Some tried to form joint ventures with local governments and still ran out of cash. The ones that survived could raise capital not only from their founders and local governments, but also from corporate venture capital. Look at the more successful EV startups, and you’ll see many backed by Chinese internet and tech giants like Alibaba, Meituan, and Tencent, as well as corporate VCs from emerging battery makers like CATL (宁德时代), the largest battery maker. With 50 or 60 startups competing, that created enormous waste and overcapacity.

Around 2020, China tried to absorb the overcapacity. The NDRC halted approvals for new EV manufacturing capacity around 2020–2021. But then came the fascinating Hefei model (合肥模式). A few months ago, there was a New York Times article about it.

Here’s what happened. NIO had some market success in 2018 and 2019, among the first Chinese startups to release EVs competing with Tesla. But from late 2019, they began running out of cash, partly because they were building so much — including many battery-swapping stations — and their NASDAQ stock price plummeted during COVID and for other reasons. So from late 2019 they started raising capital from local governments. Some of NIO’s early colleagues — colleagues of Li Bin (李斌), known in English as William Li — told me NIO was initially very reluctant, because local-government capital comes with conditions: invest a certain amount in my locality, and so on. But by late 2019 they had no choice.

Their initial supporter was, again, Beijing — but at the last minute, Beijing canceled the deal, again because of BAIC, its own SOE, which didn’t want a new competitor. NIO was desperate by early 2020. The savior turned out to be Hefei, which already hosted NIO’s factory. Hefei injected about 7 billion Chinese yuan — roughly US$1 billion — as equity into NIO. Initially, it was a win-win: NIO got the cash it needed to survive, and Hefei sold all that equity on the stock market about a year later for a three- or four-fold return — pretty remarkable for such a large investment.

After that, the Hefei model was celebrated across China as a great example of local governments acting like venture capitalists — even though they aren’t — to save a startup and boost the local economy. Ironically, when the central government tried to minimize overcapacity, the Hefei model fueled more of it. From around 2020, many cities that had missed out on EV startups rushed in, looking for new startups to back with equity. In 2020, Guangdong Province and the city of Guangzhou injected about 2 billion yuan into XPeng, adding ammunition to one of those startups. GAC (广汽) created its own startup, Aion (埃安) — you see those cars everywhere as Uber, Didi, and taxi vehicles — and it got massive equity investment from Guangdong Province and Guangzhou. Ningbo, which had missed the earlier startups, decided to collaborate with its home carmaker Geely and injected about 10 billion yuan — a massive amount — into Geely’s own startup, Zeekr.

They all hoped to replicate Hefei’s success: back a startup, then sell the stock for a two-, three-, or four-fold return. But it didn’t pan out, because the valuations of both EVs and Chinese stocks plummeted in the US after 2021. By then, though, they’d already fueled the overcapacity.

Irene Zhang: This is a good place to dig into the overcapacity. When did people in China start hearing the alarm bells, and how does it spill over?

Fengming Lu: Starting around 2023 and 2024 there were quite a lot of alarms. People are talking about it this year, especially, because you’ve seen a 20 percent year-on-year drop in China’s new-car sales, which is pretty alarming. But early signs showed up in 2023 and 2024.

The central government essentially stopped approving new automotive manufacturing plants — even EV plants — around 2022–2023. There were also new restrictions, from 2024, on local governments setting up their own private-equity investment funds. During my fieldwork in 2023, I saw many counties establishing their own investment funds, looking for tech firms — not necessarily EV firms — to invest in their localities in exchange for new business and investment.

Overcapacity has been a big issue, and people have been thinking about further consolidation among EV makers, which has proven hard. Some companies went bankrupt — most prominently HiPhi (高合) in 2023, and more recently Neta (哪吒), whose parent company is Hozon (合众新能源), last year. They were failures among the mass of Hefei-model imitators. Neta had received massive equity investment from Yichun, a city in Jiangxi Province — not a particularly wealthy place. Yichun had a lot of lithium reserves and feared missing the last-chance saloon, so it wanted its own EV maker. When Neta went bankrupt, it meant at least several billion yuan in losses for the local government’s public finances. Dealing with all this overcapacity — gradually absorbing the losses and the bankrupt carmakers — has been very challenging.

Another challenge that people outside China tend to overlook is what to do with the established SOEs. Many haven’t been very successful in EV manufacturing or engineering. Can Beijing — the central government — simply tell them to merge with one another? Since early last year, there have not just been rumours but concrete moves to have Changan, which is more successful than other SOEs, merge with Dongfeng. It took almost 12 months, and nothing really happened, because neither side’s home city — Chongqing and Wuhan, respectively — was willing to make concessions. That’s now been shelved.

Irene Zhang: Earlier, we touched on your previous work on elite politics — you have some really interesting papers about the black box at the top. How has it been doing this more down-to-earth work on EVs compared with that?

Fengming Lu: There are still several links between them. When you talk about elite politics in China, almost all senior officials have rich experience in provinces and cities. Many have connections, but they also need to show deliverables to their bosses — to prove they can get something done. For many of them, having a major automaker in their city or province is a big achievement they can sell to superiors.

Take Xi Jinping’s time in Fujian Province. Before the 1990s, there was almost no automotive manufacturing in Fujian, so one of his highlights was establishing a company called Southeast Motor (东南汽车) with a creative solution. In the 1990s there was a set quota for joint ventures — no new ones beyond the three bigs and three smalls. So Fujian’s provincial automotive company found its joint-venture partner not abroad but in Taiwan, from a company that was simultaneously manufacturing Mitsubishi cars there. They told the regulators in Beijing that it wasn’t a joint venture with a foreign carmaker but with a Chinese carmaker in Taiwan — the partner was even called China Motor Corporation (中华汽车). So it was effectively a cross-provincial joint venture: “Taiwan is one of our provinces, right? You should approve it.” That was one of the highlights of Xi’s record.

Or take Li Qiang, the current premier. His biggest achievement in Shanghai was simply Tesla — building that Gigafactory in roughly 14 months. So there’s a fascinating link between elite politics, political selection, and the auto sector: in many ways, your political performance is automotive manufacturing.

The second thing is that you always find these fascinating little stories where you need your personal connections to make things happen. Let me tell you a couple from the auto sector.

We talked about the three bigs and three smalls. The quota was supposed to be only two or three state-approved makers of passenger vehicles, because the main concern back then was that China was so poor in the 1980s — few people could afford cars, so the government shouldn’t approve too many makers. But several SOEs tried to squeeze in. FAW and the Second Automotive Works, Dongfeng, had essentially been truck makers. FAW made the Hongqi (红旗 Hóngqí), but only in tiny numbers — limousines for a handful of leaders. So they wanted approval to make passenger vehicles. The Second Automotive Works — later Dongfeng — squeezed in with the help of a good friend named Bo Xicheng (薄熙成). From the name you can already tell who he is: the son of Bo Yibo and the younger brother of Bo Xilai. Bo Yibo was extremely powerful in the 1980s, so that connection was crucial.

Then, regarding why the government in 2014–2015 opened the gate to new EV startups — having previously thought EV-making should be a game only for established automakers and especially SOEs — one big reason was a gentleman named Chen Qingtai (陈清泰), very influential among think tanks and policymakers. He wrote many policy reports arguing that China should allow more Tesla-like firms to emerge in EV manufacturing. He believed companies like Tesla represent a new, revolutionary path: you raise money from the capital market, especially venture capitalists, who do a much better job of incentivizing innovation, and you offer employees stock options to spur innovative ideas. China should copy the Tesla model.

Two things about Chen Qingtai are striking. First, he was the general manager of Dongfeng — the very company that sneaked into the joint-venture game with the help of Bo Yibo’s son. Second, he had special connections from studying automotive engineering at Tsinghua University in the 1960s, where he was a member of the university’s dance troupe. There, his “little brother” — by courtesy — was Hu Jintao. Chen was a close friend of Hu’s, and in fact sponsored Hu Jintao’s Communist Party membership. With that special connection, his policy proposals carried a lot of weight in the early 2010s. He probably wasn’t the only person who made it possible, but it was an important move that shaped the Chinese EV sector into what it is today.

Jordan Schneider: Who’d have known? This comes back to the theme. There’s an aspect of the story rich in entrepreneurial dynamism, but it’s overlaid on a system not entirely set up for it. To compete in the game of EVs and have any chance, you need to be an incredible entrepreneur — an amazing company, the best cars, product–market fit — and you need the guanxi: a friend who’s a premier or a provincial governor.

It’s fascinating that there were just enough people with enough guanxi to spin this up, so that you still get creative destruction despite the obstacles. You don’t necessarily have that in other markets, where you have to run through all these weird regulatory loopholes set up to prevent you from even starting.

Fengming Lu: In the US, you don’t really need to deal with the government until your market cap reaches, say, 10 billion or even 100 billion dollars. In China, from the beginning — if you’re a small business it’s probably fine, but if you’re starting a decent-sized business — you need to know how to deal with the government first. Maybe starting with counties or townships, then cities and provinces, and later the central government. You learn that very early on. If Elon Musk were starting his business in China, he’d probably be learning it from day one, rather than only this year.

Jordan Schneider: Totally. It’s an actual core competency you have to develop, and it doesn’t happen by accident. These CEOs are very strategic about who they want on their cap table, how much government money to take, what type to take and not take, and which politicians to bet on advancing in the future.

It’s a fascinating game. You wouldn’t necessarily think — okay, Xiaomi can design a phone, so maybe they can design an EV too; there are core competencies that align. But being good at the local, provincial, and central-government game doesn’t seem like a natural fit with being good at manufacturing cars. The fact of the matter is, you don’t get a shot at manufacturing cars unless you can figure out the political dynamics first.

Fengming Lu: Absolutely. It’s acquired to some extent, but you also need some gift or talent for it.

Irene Zhang: I’m specifically interested in whether you think the international market is the solution to all this overcapacity, or whether you have reservations about that thesis.

Fengming Lu: That goes beyond my current papers — maybe even beyond my book — and political scientists usually don’t make predictions. But I’m willing to share some observations.

Starting from my paper, what’s interesting is that Chinese carmakers are trying to copy their experience of dealing with local and central governments abroad. Take BYD’s rail-transport division: BYD is now building rail transport in Brazil — I think they’ve already delivered a project there. They see it as a way to build their public image, not only with governments but with ordinary people, signaling that they’re a good corporate citizen. Many other carmakers wouldn’t do that — General Motors and Ford never built rail transport in Brazil.

But the strategy can be risky, because many of the countries they’re investing in still hold elections. In Hungary, BYD seemed to be betting on Viktor Orbán staying in power for another term, but now they’re hitting some pushback. With a new prime minister in Hungary, their earlier plans for an automotive plant — and their hope of using Hungary as a kind of outpost in the European Union — look a bit shaky.

The second thing — and part of why people are interested in my research — is that this year’s surge in Chinese car exports is massive. By many projections, if the pace continues, China’s car exports could reach about 12 million by the end of this year. For perspective, the peak of Japanese car exports in the 1980s — around 1986, during the trade disputes between the US, Japan, and Europe — was about 6 to 7 million cars. So China is exporting almost twice as many cars as Japan did in the 1980s, when that triggered all those trade disputes.

That’s why you’re seeing new policy moves, especially from the European Union, which still makes a lot of cars — Germany, France, and so on. They’re trying to fix the loopholes; there are proposals to tariff Chinese plug-in hybrids, where before the high tariffs applied only to Chinese battery-electric vehicles. There’s also talk of tariffs similar to those imposed by the Trump administration. Over the long term, this massive export probably isn’t sustainable — you can’t imagine China exporting 15 million cars to the rest of the world next year; that far exceeds everyone’s expectations.

So China’s carmakers will need to think seriously about shifting toward more localized manufacturing — as American, European, Japanese, and Korean carmakers already have. Some are already building plants in Europe, South America, and Southeast Asia, and increasingly in Europe they’re acquiring idle manufacturing capacity, because European carmakers are also suffering from overcapacity — partly from the influx of Chinese cars (EVs and gasoline cars alike), and partly because during COVID many of them, especially in Europe, overestimated market growth. People were waiting months for new cars in Europe and the US, so automakers expanded capacity that’s now a huge burden.

It’s not really a problem for the US at the moment, but in Europe younger people are less willing to buy cars or even get driver’s licenses. Over time, the European market may suffer not just stagnation but decline. In that sense, Chinese automakers need new ways to localize manufacturing — and eventually parts supply, design, and engineering too. Many have already made some moves.

Jordan Schneider: How do you see this evolving? The story of the past 30 years is more about companies and provinces driving it than Beijing. Now that it’s a big source of international tension — a flagship industry — how do you see the relationship between the central government, the industry, and these companies changing in the coming years?

Fengming Lu: At least for the auto sector, it will still largely rest on partnerships between cities and firms. The central government already sees the EV sector as a relative success — still troubled by overcapacity, but clearly one of China’s tech successes. And over Chinese history, the central government has never really had experience directly managing the automotive industry.

There were earlier efforts. In the 1980s, the government considered creating a national-level conglomerate to manage all the automotive plants — they set up something called the China National Automotive Industry Corporation (中汽公司). But it was very short-lived, lasting only a few years, because no local government was willing to give up its carmakers. The central government has been fairly successful over the last 10 to 15 years at merging other SOEs, but in the auto sector it’s been much more challenging.

It’s a little beyond my research, but many people are interested in how this model might evolve in other tech sectors — robotics, semiconductors, AI. I’m not a semiconductor expert, but before 2017–2018 that sector followed a path very similar to autos and EVs: most new startups were products of collaboration between local governments and private firms or private capital. But because of the trade war, Huawei, and other events from 2017–2018, semiconductors suddenly became a chokepoint issue, and the central government stepped in with central-level funds.

Robotics is closer to EVs, I’d say, because it isn’t yet seen as strategically important as semiconductors — not as much an infrastructure-type thing. For AI, let’s keep an eye on it. There’s a lot of local-government involvement. DeepSeek, for instance, was essentially initiated by a hedge fund in Hangzhou and was, to some extent, protected by Hangzhou.

Jordan Schneider: That point was really interesting. Over the 2000s and 2010s, the central government wasn’t particularly excited about a hedge-fund industry existing — you had all these statements about financial speculation not creating value. But Hangzhou said, “No, this will create some rich people, we’ll get tax benefits, whatever — you can have your hedge-fund village.” Very Tang Dynasty: you hang out here by the nice streams, make some money, and we’ll just tax you. Then you’re lucky enough to have a billionaire who thinks it’s cool to train models, and DeepSeek comes out of that. It’s totally not centrally planned, but because you had regional variation that let the industry flourish, you ended up contributing to national goals in a roundabout way.

A model of the fund village in 2015. Source.

Yuhuang Shannan Fund Village (玉皇山南基金小镇). Source.

Fengming Lu: That’s true. A recent development with DeepSeek: they just did a kind of pre-IPO round and raised a lot of money. If you look at the investors, they’re dramatically different from the earlier ones. You still see plenty of private investors, VC or PE funds, and corporate VCs — but now you also see investors with a state background getting in, and some with a central-government background. That isn’t happening massively in the EV sector, because EVs aren’t quite as infrastructural, but it may happen with some of China’s key AI companies. Robotics, we’ll see. Over the next few years, let’s keep an eye on it.

Jordan Schneider: One of my favorite things to do is watch Bilibili reviews of Chinese EVs. The long list of random features these companies throw in to differentiate themselves is a font of endless fascination. Fengming, do you have any favorites?

Fengming Lu: The far-sensing is pretty innovative — I’ve seen it on some of the smart-brand cars, the brand co-engineered by Mercedes-Benz and Geely. As a traditional car guy, what I appreciate most about these cars is something simpler. Every Chinese EV I’ve driven has a genuinely good built-in navigation system. With most ICE vehicles I’ve owned, the built-in navigation exists, but you never use it — you just use your phone, because phones are far better. But Chinese EVs have very good built-in navigation in their infotainment systems, which has solved a lot of hassles for me.

Another thing I like is that in many other markets you have to pay a lot — buy the fully loaded version — to get features like ventilated or even massage seats. In China, they’re offered on very affordable models. That’s partly thanks to this evolution, and partly because the rise of the EV startups brought many Chinese parts suppliers into the supply chain. They used to be peripheral, but they became capable of building good-quality parts — seats, interiors, and more advanced ones like automotive semiconductors, braking systems, and drivetrains.

This also means a lot is changing for automotive engineers in China. I regularly do Chinese-language podcasts with engineers based in or dealing with China. One of their biggest complaints is having to meet unreasonable demands from project managers — “Add this and that within two or three weeks” — which is often impossible.

A big issue for Chinese carmakers now is how they treat their parts suppliers. Many companies have become notorious for delayed payments — some take a year to pay suppliers, essentially borrowing at their suppliers’ expense. That’s not sustainable. If the Chinese EV industry wants to prosper over the next few years, there needs to be a better arrangement between parts suppliers and the OEMs — the original equipment manufacturers.

Irene Zhang: This is a very basic point, but I can’t get over how quiet they are. I’m reporting live from Guangzhou, which is still kind of a motorbike city, and the traffic in the old parts is still chaotic. I’ve literally become bad at crossing the road because of how quiet EVs are. When you cross, you expect to hear cars approaching, but they just sneak up on you. It’s amazing.

Fengming Lu: Yes — so all those urban-noise problems got solved in a very unexpected way.

Irene Zhang: And the emissions problems too.

Jordan Schneider: You mentioned Chinese podcasts. On your homepage, it says you’ve been listening to podcasts since you had to download them to your iPod — a memory I fondly recall. Talk a little about your perspective on the Chinese podcast ecosystem. How have you seen it evolve over time?

Fengming Lu: It’s amazing. The first podcast I listened to was called Antiwave (反波), back around 2004. I listened to it not even on an iPod but on my own personal computer. There was very early podcasting in China then. You probably know him — let me find his English name —

Jordan Schneider: Goldkorn.

Fengming Lu: Jeremy Goldkorn, yes. He had his own podcast in the 2000s, and I was one of those early listeners. Later, with the arrival of iPods, I bought several iPod Classics, because they were large enough to hold all those podcasts. From around 2008 to 2010, people were downloading from iTunes — and I still remember how terrible iTunes was for downloading podcasts, full of bugs.

Most early Chinese-language podcasts were just for entertainment and fun. But starting in the 2010s, people began to realize that many US podcasts — especially the Silicon Valley type — had real, concrete knowledge in them. So Chinese shows emerged that discussed books and specific research, especially in the humanities and social sciences, from the mid-2010s — podcasts like Hu Zuo Hu You (忽左忽右). I was an early listener to those too.

From around 2020 you got a much more diverse set of podcasts. The one I take part in was created in 2021–2022. It’s called The Isle of Auto (孤岛车谈), started by a group of automotive engineers based in the UK — hence the name. It’s a Chinese-language podcast about the auto sector, and I’ve made many friends in the industry through it. It’s a relatively small show by listenership.

There are far more popular shows, many clearly trying to mimic successful American podcasts like Money Talks or The Money Show. Some have half a million, one million, or two million listeners — a lot.

There’s a big difference in how Chinese-language podcasts are consumed. In the US, most people listen through a podcast app or Spotify, with little interaction. But in China, since the 2010s, platforms like Lizhi (荔枝), and later Ximalaya (喜马拉雅), and more recently Xiaoyuzhou (小宇宙) — XYZ — built features that Chinese users appreciate. You can comment under a specific episode, like specific episodes, and discuss with people in the comment section, creating much more interaction. A lot of US podcasters probably wouldn’t appreciate those features, but they’re very successful in China.

One big problem for Chinese podcasts is that it’s very hard to get people to pay. In the US, you can run advertisements — including programmatic, customized ads, like on YouTube. But none of the Chinese platforms are really willing to try those monetization methods.

Jordan Schneider: Do you have any idea why?

Fengming Lu: I’ve discussed it with many friends. At the end of the day, most Chinese podcast listeners still prefer free episodes. A few are willing to pay for specific episodes — maybe a few thousand people — but that doesn’t make much money.

And it seems that Chinese listeners are more willing to accept what they call “soft ads” (软广) — promotional content built into the episode — and less willing to accept ad-network placements. Similar things are happening with Chinese video sites like Bilibili: whenever Bilibili tries anything with an ad network, it causes a lot of resistance from users. So even though Bilibili has so many users, it isn’t making much money, because it can’t simply use an ad network.

Jordan Schneider: It’s interesting — you see far more creator-read ads, or a food creator visiting a city because the city paid them to. There’s a lot more sponsored content than straight YouTube-style ads that have nothing to do with the show. Honestly, I’m jealous of Chinese content creators. When I post something on Spotify, I get maybe two comments, but you see these long discussions where people argue about the points the hosts make on the Chinese platforms. I hope folks learn from that.