XPeng's Path to Growth: Exclusive Insights from CEO He Xiaopeng
A revealing interview on the company's strategy for growth, autonomous driving, and the future of the automotive industry.
Ahead of the Shanghai Auto Show, XPeng unveiled their new vehicle architecture SEPA 2.0 In a post-launch interview with He Xiaopeng (Co-founder & CEO of XPeng), he discussed a series of wide ranging topics with various local media outlet. I personally don’t agree with everything that he says here, but it’s definitely a very candid discussion and worth reading.
Below are the key highlights:
Last year, we ranked 10th (in China BEV sales), but we know we can do better. We only had a few car models, and we were still building our factories, supply chains, and after-sales service and charging network. Our goal is to improve on the basics and create more than 10 models by 2025 using our advanced intelligent technology.
About assisted driving, we want to look at the big picture and see how we can turn it into reality. We've always focused on R&D, even when we were losing money. 2 years from now, we believe that autonomous driving will be a top priority for car buyers.
We'll be using the new SEPA 2.0 architecture for all XPeng vehicles through 2025, which will help us reduce cost and improve user experience. We want to run fast without spending too much, thanks to our integrated approach.
We'll still work hard on autonomous driving and even flying cars, but we're focusing more on standardization and cost optimization. We need to achieve a large scale, with millions of vehicles on the road eventually.
The toughest time for us was in 2019 and early 2020 during the pandemic. I personally contributed half of the $400 million we raised. If XPeng didn't make it, I'd pay everyone's salary and suppliers first. Last year was tough, but we had enough resources to get through it.
Now I'm overseeing all six R&D teams, so there's no doubt I'm involved in the products. Some features from our current tech will be added to the G6 through OTA updates, while others will be fully integrated from the start.
We hope to break even by 2025, but if we need to scale up even more, we might push that date back. Large-scale production is our top priority, and we want the SEPA 2.0 architecture to bring standardized intelligent software capabilities to vehicles in different price points.
In the highway NGP era, there's not a huge gap in capabilities. We're focusing on low-cost, fully autonomous NGP, and we believe it'll make a big difference. We're aiming for point-to-point semi-autonomous driving in most cities by first half of next year and even more significant improvements in the following years, marking the second critical development phase for autonomous driving. We want our products to be useful and standardized, unlike in the past when they were somewhat useful but pricey. We think the era of software changing the auto industry is almost here.
I've realized the importance of thinking big and scaling up. Annual revenue of one to two hundred billion RMB might be okay in other industries, but it's not enough for the automotive industry. I've taken charge of many tasks myself and thought deeply about various products and ideas.
As for my previous comment producing 3 million vehicles per year in order to survive, I think for us it should be closer to 5 million. If you're a mid-to-high-end brand, 3 million could be possible, but it's the minimum threshold.
To excel in autonomous driving, you need persistence and a well-structured organization, not just a few strong individuals. XPeng's organizational structure wasn't great in the past, and I've learned that neglecting structure can lead to significant losses. After noticing some major issues with XPeng, I made significant adjustments. Now, with Wang Fengying on board, she is helping us improve our product planning process and efficiency. We're also learning from successful companies like Toyota and Volkswagen on best practices for cost control and org structure.
We're far ahead in technology and market, but cost is a crucial capability. We'll launch the NGP with XNet in June, and the experience will be completely different.
800V isn't that difficult. I'm personally in charge of it, and I don't know why people keep asking about it. The key is getting the battery cells and vehicle right.
Our semi-autonomous driving capabilities are leading in China. We're also exploring integrated die casting internally, but when considering the full lifecycle cost of die casting, just replacing the parts is both faster and cheaper.
I'm confident comparing our XNGP compared to Tesla's FSD. Our experience in China this year will surpass FSD's in the US. However, cost is more important, so we'll wait until next year to compare with Tesla on that front.
I’m not sure who will be part of the global top 8 brands. But I believe to become part of this elite group requires user-oriented innovation, global-scale sales & delivery network, and organizational management capabilities to integrate hardware and software together.
Scaling up can reduce costs (from economies of scale), but innovation is needed for further reductions. With scale, maybe it’s possible to go from 100 RMB / part to 70 or 50 RMB / part. However, to reduce from 50 RMB to 10 RMB / part, innovation is definitely needed.
In traditional automotive industry, new auto startups can't really compete with Volkswagen and Toyota. EV is also hard, but breakthrough is possible by combining the transformational power of battery technology and software. Using China's positioning method globally would cause problems.
Tesla wasn't the first to develop assisted driving, but they achieved scale, which is impressive. We started focusing on it in 2014, and although we have some overlap with Tesla, our logic is mostly different. Today, all companies work based on Google's Transformer paper.
(Original content was posted by 不是郑小康 on Weibo, then translated and summarized by me)