After eight years, Pony.ai, once the most highly valued autonomous driving company in China, finally went public.
Pony.ai’s entrepreneurial story began in late 2016, a prime year for autonomous driving.
Flush with cash, internet giants entered the fray with confidence. Uber acquired the self-driving truck company Otto for USD 700 million, while Google’s autonomous driving project Waymo became an independent entity. Traditional automakers, eager for transformation, also jumped in with their self-driving plans. Meanwhile, chipmakers like Intel partnered with BMW and Mobileye to enter the space.
At Baidu’s autonomous driving division, James Peng and Lou Tiancheng decided over a cup of coffee to leave their stable positions at a major corporation to establish Pony.ai. During the years when the mobile internet boom had peaked, autonomous driving became an industry receiving the most investor attention. Naturally, the combination of these two renowned technologists attracted top-tier institutional investors, and Pony.ai’s valuation soared, solidifying its position as a leader in the autonomous driving field.
In many ways, Pony.ai’s fundraising history is like a smaller version of Nio’s story. From being surrounded by top-tier USD venture capital funds, to embracing industrial funds from traditional automakers, and finally attracting sovereign wealth funds from the Middle East, Pony.ai was fortunate enough to survive the different stages of investors’ honeymoon with autonomous driving.
In June 2021, with the help of financial investors, Pony.ai began its IPO process. It was supposed to go public at a time when both capital and confidence were at their peak. However, as the climate for Chinese companies listing in the US worsened, Pony.ai’s IPO had to be suspended.
Today, autonomous driving is no longer considered an appealing investment. Massive losses are detailed in Pony.ai’s prospectus, and its market capitalization has been halved compared to its last funding round.
For the later-stage sovereign or industrial funds that bet on Pony.ai, as well as the early financial investors who once anticipated a USD 10 billion IPO three years ago, the current low-hanging fruit is clearly less satisfying than they hoped.
This, perhaps, is the stark reality that technology investors must face: If a company does not seize the moment when the technological winds are in its favor, the bright young talents in the commercial world will not be widely embraced by the broader market.
Through interviews with nearly ten Pony.ai investors and the company’s founder, Peng, 36Kr attempts to explore the question: When a new technology company faces valuation declines and real commercialization challenges, when is the best time for venture capital to invest?
The final stretch
In June 2021, Pony.ai was just one step away from its IPO.
Four months earlier, the Chinese autonomous driving company completed a USD 250 million Series C+ funding round at a valuation exceeding USD 5 billion. Investors included the Ontario Teachers’ Pension Plan (OTPP) through its Teachers’ Venture Growth (TVG) program (formerly Teachers’ Innovation Platform), the Brunei Investment Agency, and CPE.
Sources familiar with the financing revealed that this occurred when Tesla shares were surging in the secondary market and Waymo’s fundraising was smooth. Large foreign pension or sovereign funds—many of which serve as limited partners for Chinese general partners—were opening offices in mainland China or Hong Kong. Across the ocean, the highly valued Pony.ai naturally caught their attention.
By then, Pony.ai had accelerated its expansion. In December 2020, Pony.ai established a trucking division and registered PonyTron, which at its peak had a team of around 100. By June 2021, it had set up a vehicle development team of over ten people in Shanghai’s Jiading district, marking its entry into vehicle manufacturing.
All actions pointed toward going public. An investor said that an IPO was seriously being considered at that time, with rumors at one point alluding to a valuation that exceeds USD 10 billion, or even USD 12 billion.
But a black swan event changed everything. In August 2021, the US Securities and Exchange Commission (SEC) called for a halt on IPOs for Chinese companies. After more than a year of preparation, Pony.ai’s listing was canceled due to force majeure. That autumn also marked the beginning of the decline for USD funds in China.
“A fast-moving vehicle that suddenly encounters obstacles will inevitably face turbulence,” one investor said. First, the vehicle manufacturing plans were indefinitely delayed. Several core employees from PonyTron left to start their own companies, one of which was even funded by Pony.ai’s old investors.
The turbulent period continued. In October, a Pony.ai test vehicle hit a road sign while changing lanes in California, leading to the first-ever recall of a Level 4 autonomous driving system.
Pony.ai then considered listing in Hong Kong or mainland China. “There were said to be ongoing discussions about potential opportunities, but Pony.ai had yet to achieve profitability.
Peng admitted in an interview that the team faced a significant challenge due to the industry’s uncertainty—large-scale commercialization was not yet visible. He raised the question of whether the company would have enough resources to sustain development if commercialization dragged on longer.
An investor remarked that, while the industry remained obscure in 2017 and 2018, by 2021—similar to the present situation with large language models (LLMs)—the market became more transparent. Challenges in commercialization or technology could be discerned through due diligence. Combined with wavering confidence in Chinese stocks, most believed that if Pony.ai did not go public, it would struggle to raise money in the primary market at its USD 5 billion valuation.
No one knew how long the winter would last. One investor recalled Pony.ai jokingly telling them about regrets of not raising more funds earlier.
Apart from staff reductions, business downsizing, and a pivot to products that could generate immediate commercial returns, Pony.ai also sought alternative paths.
When Level 4 autonomous driving faced hurdles in complex, low-probability scenarios, the company turned to Level 2 advanced driving assistance systems as a temporary lifeline. At one point, Pony.ai even aimed to position itself as a supplier of driving assistance technology.
After the failed IPO attempt, Pony.ai allocated some resources to develop and enhance Level 2 driving assistance systems. In Guangzhou’s Nansha district and Beijing’s Yizhuang district, Pony.ai vehicles without light detection and ranging (LiDAR) were frequently seen on the streets, offering driving assistance.
Amid these challenges, Pony.ai managed to stay afloat.
Racing for investments
Let us rewind to 2016, the dreamlike beginnings of Pony.ai.
After deciding to start the company, Peng’s first call was to Zhou Kui, a partner at HongShan (formerly Sequoia Capital China).
Earlier that year, Zhou attended Sequoia’s annual meeting in Silicon Valley, where he met Peng at Baidu’s US research lab. In October, Zhou received a call from Peng, who had just returned to Beijing and was considering partnering with Lou to start a self-driving venture. Zhou quickly brought in Neil Shen and Fu Xin, also from HongShan, to meet Peng.
Lou recalled a preinvestment dinner with Shen, during which he was posed numerous unrelated and seemingly random questions, none of which had to do with autonomous driving. Lou answered them one by one. Later, Shen explained that he wanted to assess Lou’s overall personality through casual conversation.
During its Series A financing, Pony.ai opened up 20% of its shares. HongShan led the round, investing USD 10.5 million for a 15% stake at a USD 70 million valuation. HongShan continued to invest in Pony.ai across four rounds and is now the company’s largest financial investor with a 10.2% stake. Due to the relevance of Pony.ai to multiple tech and automotive sectors, three HongShan partners were involved in the investment, which is rare.
“2016 was still the early days for autonomous driving,” Zhou told 36Kr. “Our judgment at HongShan was that the field had immense potential but would require persistence and long-term investment. There are no shortcuts.” He also emphasized that it was a field requiring highly talented technical founders to gather in one place to innovate.
Peng was one of Baidu’s earliest US researchers and served as the chief architect for Baidu’s self-driving efforts. He was one of the very few engineers at Baidu to hold a T11-level title. Lou, known as one of the world’s top programmers, had worked at Google’s self-driving team and was Baidu’s youngest T10 engineer. Early investors believed that such a founding team would naturally attract other talented technologists and lay the foundation for autonomous driving.
One investor shared an anecdote: whenever Lou hosted a dinner, alumni from Tsinghua University’s prestigious Yao Class would join without hesitation.
According to Peng, the first round of financing did not involve massive roadshows or excessive focus on valuation. The founders only reached out to four investors, and HongShan and IDG Capital chose to support the company. During Pony.ai’s second financing round, HongShan facilitated Nio Capital’s investment in the company.
Another key early investor was 5Y Capital (formerly Five Source Capital). In 2017, Zhang Fei, a partner at 5Y Capital, visited Pony.ai’s Beijing office for the first time and met Peng. Peng described the meeting as a “moment of mutual understanding.” Zhang’s previous investment in Kuaishou became a significant reason for further discussions between 5Y Capital and Pony.ai.
In its early days, Kuaishou had only one founder, Cheng Yixiao, and 5Y Capital invested RMB 2 million (USD 280,000). A year later, when Kuaishou hit a bottleneck, Zhang persuaded Su Hua to join the company by proposing a solution: both 5Y Capital and Cheng’s team would dilute half their shares to create an options pool, with most of it allocated to Su’s team.
This was one of those rare moments where an investor altered the trajectory of a company.
Zhang told 36Kr that Peng and Lou both possessed strong engineering skills and were highly pragmatic. When investing in entrepreneurs, he often worried about idealists being disconnected from reality. “But the style of these two founders is remarkably balanced,” he said. Zhang believed that, in a field like autonomous driving, which requires a complete redefinition of its principles, “the only thing you can truly rely on is the team.”
5Y Capital’s lead investment in the Series B round was the one that truly put Pony.ai on the map in the investment community. By then, Pony.ai had already developed demonstration vehicles, and its team had grown to over 20 people. “It was an excellent fundraising round,” Peng recounted, adding that his schedule was packed, meeting at least five or six investors every day.
Shen Xiao, a partner at Eight Roads, recalled meeting Peng and Lou at a hotel in the US. “There wasn’t much preamble—they opened a whiteboard right away and began discussing Level 4 technical details,” Shen said. “At that time, Level 4 was still quite distant—LiDAR and drive-by-wire redundancy were both far from mature.”
Yet, Shen Xiao told 36Kr convinced that Level 4 autonomous driving could be achieved and that it could be simplified to target Level 2 use cases. After a second meeting, Eight Roads decided to invest.
Due to overwhelming investor enthusiasm, Pony.ai’s valuation was adjusted over three times during this round. The Series B+ round began almost simultaneously, reaching a valuation of USD 900 million. “From July to September, sensor solutions were being iterated constantly. The market was strong, and everyone accepted these valuation jumps,” Shen Xiao said.
Lin Haizhu, founding partner of Jardine Asia, noted that, before 2020, Pony.ai maintained a rhythm of raising one round every six months. When Lin engaged with the team, the Series B round had already closed. For the Series B+ round alone, over 40 term sheets were submitted, with the final selection narrowing to around ten funds. Lin, being an alumnus of the same school as Peng and Lou, leveraged that connection to join the round.
Also participating in this round was Gong Biao, managing director at VMS Group, who told 36Kr that the market had very high expectations for autonomous driving. “There was a sense of competition [among investors] to secure shares—better to make a mistake than miss out,” Gong said.
A brain needs a body
Toyota’s involvement in 2020 marked a pivotal moment in Pony.ai’s fundraising history. For intelligent driving companies lacking vehicle manufacturing capabilities, partnering with automakers is essential. If autonomous driving is likened to a brain, then cars are the body.
“No matter how brilliant a brain is, it needs a healthy body,” said Lou in an interview.
In February that year, Pony.ai raised USD 400 million from Toyota at a valuation of USD 3 billion—double its valuation from the previous round. The two companies’ collaboration dates back to August 2019, when Pony.ai began testing on public roads in Beijing and Shanghai using Lexus RX vehicles, a Toyota subsidiary.
The financing deal with Toyota was the culmination of six or seven trips Peng made to Japan over a span of half a year. Negotiations progressed gradually, moving from specific project collaborations to strategic partnerships, and finally to investment. “There was no pivotal moment or single decisive event; it was a process of slowly heating the water until it boiled,” Peng said.
Toyota told 36Kr that its primary motivation for investing in Pony.ai stemmed from its belief that China will remain the world’s largest market for highly autonomous mobility services over the next 5–10 years and beyond. To accelerate the commercial deployment of robotaxi services, Toyota chose to collaborate with a local technology firm.
Why did Pony.ai win Toyota’s favor? According to one investor interviewed by 36Kr, another leading autonomous driving company, Roadstar, was initially closer to Toyota. However, Roadstar’s internal conflicts over business strategy and equity distribution caused the company to fall apart, with its former CTO founding another startup, DeepRoute.ai. “Pony.ai seized the opportunity,” the investor said.
Peng offered a different perspective, stating that Roadstar had never even met with Toyota executives and was far from securing investment. He noted an intriguing detail: before Xiaomi officially announced its carmaking ambitions, it had also considered investing in Pony.ai. Peng had multiple discussions with Xiaomi’s Lei Jun. However, Xiaomi ultimately chose to acquire Deepmotion—a cheaper autonomous driving solution company—for USD 500 million. “Xiaomi’s core DNA lies in creating good products with extreme cost control and high value-for-money,” a Deepmotion investor said.
Pony.ai pinned its hopes on robotaxis, which needs Level 4 autonomous driving controls. Consequently, its partnership with Toyota is better described as joint development. For instance, the fifth, sixth, and seventh generation of its autonomous driving system were first installed on Toyota’s vehicle platforms.
That said, Toyota is not solely committed to Pony.ai. It established its own autonomous driving division in 2018 and has invested in Uber’s self-driving division and Momenta in China.
The real test
“A top-tier company has many lives,” said a Pony.ai investor, encapsulating the company’s fate.
After its IPO setback, a lifeline of funding fortunately arrived in March 2022, when Pony.ai announced the completion of a USD 190 million Series D round led by the China-Arab Investment Funds, with participation from OTPP, 5Y Capital, G&O, Carlyle Group, ClearVue Partners, Raumier Limited, Evodia Investment, and Assetkey Limited.
In October 2023, Pony.ai announced an additional USD 100 million investment from Saudi Arabia’s Neom and its investment arm, the Neom Investment Fund (NIF).
Pony.ai’s valuation during the Series D round was USD 8.5 billion. Sources close to the financing revealed that Middle Eastern countries, in the context of declining USD funds and increased attention to geopolitical factors, saw this as an opportunity to acquire quality companies at a discount. “In these investments, approximately 80% of the focus was on industry relevance, while the remaining 20% considered financial returns.”
At the time, Middle Eastern capital primarily flowed into traditional energy projects, with few technology, media, and telecommunications investments. Pony.ai was among the earliest recipients of such funding. “With USD funds drying up and RMB funds having shorter exit cycles, it’s difficult to invest in projects with long commercialization paths, like autonomous driving. The Middle East is the most promising alternative,” Peng told 36Kr, adding that securing Middle Eastern investment was a deliberate strategy.
However, some investors contributed through a combination of new equity and secondary share purchases. One investor noted that, while buying in, the price of new equity was around USD 7 billion, while secondary shares were priced at approximately USD 1 billion.
Although the top-tier positioning of Pony.ai as a leader in autonomous driving helped maintain its reputation, the real challenge lies in the secondary market. In stark contrast to its widespread appeal in the primary market, Pony.ai’s Nasdaq valuation has halved from its last funding round.
Multiple investors told 36Kr that the current valuation represents a consensus price range among capital market participants.
Peng responded to the media outlet by stating that, while primary market valuations are driven by resources and external confidence, secondary market valuations depend on performance. “It’s time to take the horse to the track and see if it runs,” he said, emphasizing the need to seize the current rare IPO window.
Three years ago, when Pony.ai was preparing to go public, its valuation exceeded USD 10 billion, higher than Nio’s current market cap. Three years later, Pony.ai has obtained permits to operate autonomous driving services in Beijing, Shanghai, Guangzhou, and Shenzhen with a fleet of over 250 robotaxis. Yet, its market value is less than half of its previous peak.
Pony.ai’s prospectus reveals the stark reality: from 2022–2023, the company posted net losses of USD 148 million and USD 125 million, respectively. In the first half of 2023, it reported a net loss of USD 51.78 million. The commercialization of autonomous driving appears as elusive as the “iPhone moment” for virtual reality (VR)—perpetually anticipated, but never arriving. Secondary market investors, focused on fundamentals and valuations tied to financial metrics, are unwilling to pay for the yet-unproven robotaxi story.
Six years ago, Li Jiaqing, in explaining Legend Capital’s rationale for investing in Pony.ai, said, “The key question is whether this represents the most critical transformative direction. If it is, we invest, especially in early-stage, technology-driven ventures. It’s hard to assess with the mindset of a mature industry. Asking upfront about business models? We don’t know yet.”
Lu Huangxian, managing director of Redpoint China Ventures, echoed this sentiment, noting that the firm understood it was a long shot at the time of its investment.
Buying into technological dreams and trading time for exponential growth was once a foundational logic for venture capital. Today, as realism and focus on fundamentals dominate the narrative, late-stage investors who once flocked to unicorns must now confront micro-profits—or even losses.
If the internet funding frenzy was a flowing banquet of capital, Pony.ai exemplifies the fate of a technology company when its investors return from dreams to reality. A confluence of factors shaped its trajectory: shifting US-China relations, changing investor preferences, difficulties in fundraising, and the rise of a “DPI above all” mentality. But this was no accident—it might even become the norm.
Broader implications
The more pressing question is whether extreme realism is correct. What kind of capital environment does true innovation require? It’s worth remembering that autonomous driving emerged from a daydream.
One Pony.ai investor told 36Kr that having the right assets is more important than timing. Perhaps this is a form of self-consolation. They might still recall the story of Pinduoduo: valued at under USD 30 billion on its IPO day, it dipped below its offering price at least four times in its first year, yet now exceeds USD 130 billion in market capitalization. Perhaps Pony.ai’s public listing is just the beginning.
Over two months ago, renowned angel investor and Neu Venture Capital founder Jerry Neumann announced that he would no longer invest in new startups. In his farewell to a career as an individual venture capitalist, Neumann said that today’s startups no longer feel like they are driving progress but have become more about running a business. Perhaps, that is but a reflection of the current investment climate.