Thursday, 2022 December 8

Tech turns to the automotive industry: DJI vs. Huawei

DJI and Huawei, giants in their respective fields, have many similarities. They were both established in Shenzhen and for the time being, neither has plans to go public.

The only difference is that they’re big players in different sectors: one in global communications equipment (Huawei), and the other in commercial unmanned aerial vehicles, or drones (DJI).

Both companies recently made the decision to enter the automotive industry as all-in-one software and hardware solution providers, to help vehicle manufacturers build better, smarter models.

Their foray into the automotive market

Some time ago, SAIC-GM-Wuling Automobile announced the results of its collaboration with DJI, with the official release of the Lingxi Intelligent Driving System. The system is said to be able to “achieve any type of obstacle identification and response.”

However, unlike almost all manufacturers on the market, the Wuling model equipped with DJI’s system does not use LiDAR (light detection and ranging) sensing hardware or even an automated driving chip with high computing power. Instead, DJI capitalized on algorithms they had developed in their UAV technology to build their intelligent driving function for small vehicles, enabling them to offer a competitive price of RMB 100,000 (USD 14,000) per vehicle.

KiWi EV. Source: SAIC-GM-Wuling

In contrast, vehicles equipped with Huawei technology, such as the BAIC Arcfox, the Changan Avatar, and the Great Wall Saloon, are all higher-end models priced between RMB 300,000 to 400,000 (USD 42,000 – 56,000). Though they come with a significantly higher price tag, these models have the assurance of a strong hardware foundation, such as LiDAR, for their high-level assisted driving functions.

In this way, DJI and Huawei have adopted two different methods—the former relies on a good performance-cost ratio, while the latter increases value by delivering robust functionality.

Motives for entering the industry

Huawei’s and DJI’s motives for venturing into the vehicle industry are almost identical.

Huawei’s smartphone business suffered a huge blow in 2020 due to US sanctions. Statistics from International Data Corporation, a global market intelligence firm, show that Huawei’s global cell phone shipments in 2020 totaled 189 million units, down 21.5% year on year. As a result, Huawei’s consumer business revenue in 2020 was RMB 482.9 billion (USD 67.6 billion), only a slight year-on-year increase of 3.3%.

DJI’s story is similar. Though its global market share of the consumer UAV market was significant, it was declining due to tightening industry controls and flight bans, and the company’s revenue growth slowed. Data shows that DJI’s revenue and net profit only grew by single digits in 2018 and 2019, and the company did not achieve a year-on-year revenue growth rate of 30% and a net profit growth rate of 33% until 2020.

A DJI Air 2S drone flying. Image courtesy of Silvio Wiggelinghoff

For both of these companies, the intelligent automotive industry presents an attractive prospect, as it is still in the early stages of development and has a bright future. According to the 2022 China Intelligent Vehicle Development Trend Insight Report, by 2025 the sales volume of intelligent vehicles at L2-level and above in China is expected to exceed 10 million units, corresponding to an intelligent vehicle penetration rate of 49.3%. The size of the smart cockpit market is expected to exceed RMB 100 billion (USD 14 billion), making it a promising market.

Areas of competition

When dividing up the intelligent vehicle market, both companies chose different paths. Huawei started from the high-end sector, while DJI took the cost-effective route.

This has resulted in different types of collaborations and output from the vehicle manufacturers they have partnered with.

Those that choose to collaborate with Huawei tend to produce higher-end models with increased prices. For example, Great Wall Motor, a brand where the average price of a car was once RMB 80,000 (USD 11,200) per unit, launched a high-end model, Saloon, at a price of up to RMB 488,000 (USD 688,300). The Saloon model relies on Huawei’s LiDAR sensing hardware and computing platform to support its intelligent functions.

Another instance is Changan, which usually sells cars at an average price of RMB 100,000 (USD 14,000) per unit, which launched the high-end Avatar range in collaboration with Huawei. Avatar’s limited-edition model is RMB 600,000 (USD 84,000).

A race car equipped with Huawei autonomous driving solution. Source: Huawei

On the other hand, most of the brands in collaboration with DJI are producing middle and entry-level models that are geared towards the masses. For example, the XPeng P5, which is equipped with two HAP LiDARs from DJI’s subsidiary Livox, is priced at RMB 209,900 (USD 29,400). The second DJI-equipped model is Wuling’s Baojun KiWi EV, priced at RMB 102,800 (USD 14,000).

The influence of their current positions

The difference between the companies’ methods of entry into the automotive industry is related to their current positioning in their own sectors. Huawei has always been positioned as a maker of high-end products, with the Huawei Mate series and the P series considered flagship products in the smartphone industry. Its Huawei Smart Home Solution also far exceeds other players in the industry in price, which is as much as RMB 99,999 (USD 14,000).

In contrast, DJI’s approach in the UAV market has always been centered on aerial photography performance, quality, cost performance, and ease of use. The company launched its Phantom series at the end of 2012 in a market where UAVs were often very expensive, focusing on making UAVs accessible to the masses.

In essence, Huawei’s approach is to increase value by stacking technology and features in its products, while DJI’s approach is to ensure cost-effectiveness by using algorithms to compensate for the limitations of its hardware.

Will this cost-effective method work for DJI outside of the UAV sector? It’s tough to say with certainty, since performance and reliability are often higher priorities in the vehicle industry.

DJI’s subsidiary Livox is able to produce lower-cost systems due to using a double-wedge prism scanner in its LiDAR system, which has fewer components. However, the tradeoff is the lack of real-time information, as the system is slower. Such a system would work in static scenarios such as mapping terrains and ensuring security.

Livox’s first automotive-grade LiDAR product for serial production. Source: Livox

However, intelligent driving systems are often put in much more complex and dynamic scenarios, moving at high speeds and coping with ever-changing road conditions. These require a much faster response time and higher sensitivity from their sensors, which DJI’s double-wedge prism scanner may not have.

After experiencing setbacks with its LiDAR business, DJI pivoted and began selling all-in-one solutions, continuing down the cost-effective performance route. The company currently provides all-in-one software and hardware intelligent driving products, including D80, D80+, D130, and D130+, among which D80 and D130 are mainly basic L2-level auxiliary driving functions and include ACC (adaptive cruise control), auto-following, auto parking, and other functions. D80+ and D130+ are more advanced versions.

These solutions were also integrated into its partners’ vehicles, such as the 2023 KiWi EV model designed in collaboration with Wuling, which has a D80 system. This means that the KiWi EV model can only perform assisted driving at speeds below 80 km/h, and this function mainly covers urban areas and expressways. As one of the simplest products in the DJI lineup, the D80 does not use LiDAR or high-precision maps, and the computing power of the whole system is only 16 TOPS (trillion operations per second).

As a comparison, the Arcfox Alpha-S HI—made in collaboration with Huawei—uses the Huawei MDC (Mobile Data Center) intelligent driving computing platform, which has a computing power of 400 TOPS and contains two Nvidia Orin X chips.

This system is also used in Changan’s Avatar 11, Great Wall’s Saloon, the Li Auto L9, the XPeng G9, and the SAIC Feifan R7.

BAIC Arcfox Alpha-S HI, a special model in collaboration with Huawei. Source: BAIC

Different visions

From a tech perspective, DJI’s approach has the advantages of low costs and high scalability. The company directly transplanted its UAV technology into its intelligent vehicle systems. This technology is based on binocular vision and point cloud algorithms and is used to obtain geometric information, including depth information, to judge whether the obstacles ahead will threaten driving safety and thus eliminate the dependence on LiDAR.

However, this results in limited perception and computing power. DJI’s systems, when encountering scenarios they cannot cope with, will return control to the driver—a situation that may be dangerous, especially if the driver is distracted or assumes the system can handle the obstacle automatically. An example of this is the lack of high-precision mapping in the Lingxi Intelligent Driving System, which requires the driver to take over at misplaced intersections and roadworks.

On the other end of the spectrum, Huawei has introduced a full support system on the Arcfox Alpha-S HI version, covering sensing, decision-making, steering, braking, and power. The inclusion of such a system comes with a corresponding price tag, but it means it’s powerful enough to help the vehicle pull over during extreme circumstances and cope with a much larger range of scenarios.

Both companies have some way to go in development. DJI’s product vision may have to be adjusted if it wants to compete against other intelligent vehicle systems. Huawei, while having the advantage of having mass-produced and installed its powerful hardware, still has some improvements to make for its software upgrades.

Regardless of how both companies adjust their path forward in the automotive industry, user experience and system reliability are essential for success. In the competition for orders from vehicle manufacturers, who will come out on top depends on how they adapt in the next few crucial years.

This article was adapted based on a feature originally written by Brother Yu and published on Box Rice Finance and Economics (WeChat ID: daxiongfan). KrEurope is authorized to translate, adapt, and publish its contents.

KrASIA Connection
KrASIA Connection
KrASIA Connection features translated and adapted content published by 36Kr.com, the largest and most influential tech portal in Chinese language.
MORE FROM AUTHOR

Related Read