10,000 yuan a year, the price of agricultural drones "flies close to the ground"

Agricultural drones have gradually matured from the initial outbreak.

In late November, XAG CEO Peng Bin said in an interview that the company turned losses into profits this year. The prices of major enterprises are also continuing to be "civilian", and the products with a load of 70 or 80 kilograms have recently dropped to the 40,000 yuan mark.

Analyzing this price reduction curve, around 2016, the price of XAG drone kits was still around 100,000 yuan, and the starting price of the newly released P150 Pro dropped to less than 40,000 yuan. If you look at the load per kilogram, the unit price of DJI MG-1P load in 2017 exceeded 5,000 yuan/kg, and the T60 in 2023 dropped to around 1,000 yuan/kg.

There is a market for agricultural drones, and everyone no longer doubts it. However, the price of the machine has dropped by several thousand yuan or 10,000 yuan every year, and how much more can it be reduced?

Price reduction, but not loss for scale

According to the data of the observation and research report, industrial drones will account for nearly seventy percent of the domestic civilian drone share in 2023. In 2022, agriculture and forestry plant protection will surpass geographic surveying and mapping to become the largest application field of industrial drones, accounting for more than one-third of the total. This data may be more intuitive: according to the UAV cloud exchange system data released by the Civil Aviation Administration, the operation of plant protection UAVs accounts for more than 98% of the flight hours of all cloud system UAVs. Since the beginning of the rise of agricultural and forestry plant protection drones or agricultural drones, price competition has been an enduring topic, and it is not uncommon to exchange losses for scale, which has an impact on the profitability of manufacturers.

At the end of 2017, DJI announced a new round of price reduction measures for plant protection machines, saying that DJI Agriculture would get rid of profit-oriented goals and focus on improving industry efficiency and building a closed loop of services. According to the prospectus previously submitted, XAG's revenue from 2018 to 2020 increased from 320 million yuan to 530 million yuan, and the net loss also expanded from 6.717 million yuan to 60.846 million yuan, and the net loss in the first half of 2021 reached 85.123 million yuan.

Prices are lower than the competition. In 2017, XAG launched a P202018RTK with an operating load of 10kg and a bare metal price of 41,999 yuan, and more than a month later, DJI released the MG-1S Advanced with an operating load of 10kg, priced at only 29,999 yuan. With some market players losing out on performance or price, agricultural drones are gradually becoming a highly concentrated industry. According to the statistics of the national agricultural machinery purchase subsidy system, in 2023, a total of 35,631 plant protection drones will be subsidized in the national agricultural machinery subsidy, of which DJI will account for 80.54%. According to QY Research's 2024 sales forecast, DJI, Yamaha, XAG and Autel Robotics account for 30%, 11%, 9% and 5% of the global agricultural drone market, respectively.

Judging from the current situation, the price of small-load agricultural drones can reach more than 20,000 yuan, and the price of large-load drones per kilogram of load has also dropped to less than 1,000 yuan. While the drone load is increasing this year, the price of the P150 Pro launched by XAG in November starts at 39999 yuan, and the price is nearly 4000 yuan less than the product launched last year, and the T70 launched by DJI last week starts at 44999 yuan, and the price is 10,000 yuan less than the previous year's T60.

However, there have been recent indications that the price of products is still falling and it is not a simple price war. The ability to continue to reduce prices is related to the decline in costs. "We use drones to 'pick' the cost one by one, because the farmer can't accept such a high price, and he doesn't want defective products, we save the province, including transportation, packaging, those packaging boxes we are not so particular." Peng Bin said that just like Musk's rocket X, it is simple to normalize complex things, simple will be cheap, and the cost reduction of 70%~80% is determined by design.

DJI Agricultural Technology Solutions engineers also recently mentioned to reporters the role of product design in cost savings. "The entire cost of drones is 'cut' down, and the logic is very similar to that of electric vehicles." The above-mentioned engineer said that the UAV also has three electrics, and in the case of larger propellers, it is more difficult to mass produce, but the power efficiency of the propeller power system has been improved, so that the power and power consumption have decreased, and the battery cost has decreased. Secondly, the drone can also reduce weight, balance the strength of each link, and deduct excess weight. In addition, just as electric vehicles reduce assembly labor costs, material costs, and post-processing costs by merging parts through an integrated process, drones also reduce parts through integrated design.

In addition, there are some cost reductions that are hard to ignore come from the supply chain, and the price of drone hardware components has also decreased.

The price of key components has been significantly reduced

The declining cost of agricultural drones could be an excellent example of the interplay of the hardware industry chain. Around the world, the automotive industry and the robotics industry have become more and more closely linked in recent years. Manufacturers that have laid out new energy vehicles have begun to make robots, some robot manufacturers have begun to do autonomous driving, and even automotive chip manufacturers have extended to robot chips. Behind it, not only robots and cars belong to the same category of embodied intelligence, but there are also similarities in hardware, such as batteries, radars, and chips.

Some people in the UAV industry told reporters that in the past few years, benefiting from the development of the robot industry and the electric vehicle industry, the price of raw materials such as lithium has fallen with the expansion of applications, and the battery accounts for about 1/3 of the cost of UAVs, and the annual price reduction is about 10%.

The above-mentioned engineer also told reporters that the price reduction of lidar, which was originally mostly used in cars or other robots, is also very obvious. Lidar accounts for a high proportion of UAV perception systems, and the T100 released by DJI this year is equipped with LiDAR as standard for the first time, and the price of this LiDAR has dropped by multiple in the past few years. It is expected that LiDAR will continue to double the performance of the same price or halve the price of products with the same performance every year, and this trend has only just begun.

The T100 uses Livox's Mid-360 3D LiDAR, an independent LiDAR company incubated within DJI. Judging from the information released by other top listed lidar manufacturers, the price of lidar is rapidly entering a downward channel.

According to data from Yole Group, a research institution, the average selling price of LiDAR systems in China will be US$450~500 in 2023, a significant year-on-year decrease and lower than the selling price of US$700~1000 in other overseas regions. At the recent performance exchange meeting, Qiu Chunchao, CEO of Suteng Juchuang, said that the laser radar in the field of robotics benefits from the research and development achievements of chipization and digitalization in the direction of vehicle lidar, and under the traction of vehicle demand, the cost of high-performance lidar can even sink to the acceptable range of consumer robot applications, and the growth rate of the robot field shows explosive growth similar to that of the automotive industry. Li Yifan, CEO of Hesai Technology, recently revealed that the company plans to halve the price of its main lidar products next year.

The large-scale growth of drones in the past few years has also led to a decrease in the cost of other components. According to the above-mentioned engineer, the price of a positioning device on the drone has dropped by more than 100% since 2015. A few years ago, an electromagnetic flowmeter for industrial use was sold for eight or nine thousand yuan on the market, and now it is sold for two or three hundred yuan with the same accuracy. UAV manufacturers need to consider intelligence, reliability and make the price acceptable to users, in the case of declining component costs and new technology maturity, UAVs in recent years have gradually made new components such as electromagnetic flowmeters, lidar and other new functions "on the machine".

However, the curve of cost reduction may not be as steep as before. DJI's payload unit price curve shows that from the products released in 2017 to the products released in 2019, this unit price decline curve is quite steep, from 2019 to 2020, the curve is a little flatter, and after 2020, the curve is even more flat.

The above-mentioned engineer explained that the scale of UAVs has not fully come up in the past few years, and the scale was less than half of the current scale three years ago. After that, the price decline due to the combined effect of technological evolution and cost reduction is still smaller than the previous decline due to the effect of scale. In addition, in order to improve performance and intelligence, more sensors and other components will be superimposed, and in this case, it will be more difficult to ensure a stable price reduction, which depends on the iteration of product development and design.

Agricultural application scenarios can continue to expand

In terms of product quality, there have been many "bombing" incidents of agricultural drones in 2021, and such incidents have rarely appeared in the public eye in the past two or three years. In terms of coverage, data from the National Agricultural Technology Center show that in 2024, there will be 251,000 unmanned plant protection aircraft and 2.67 billion mu of control operations in the country, an increase of nearly 25% over last year, ranking first in the world. In terms of the area of cultivated land covered, the data provided by DJI is that in 2024, the number of agricultural drones will exceed 2.5 billion mu in a single year, which is equivalent to covering 38% of China's cultivated area.

When the agricultural drone market matures and covers a large area of arable land, and companies begin to consider sustainability and profitability, does it mean that this product will move towards a stable form and stable function, and no longer have greater imagination space?

In an interview with the first financial reporter and other media,Peng Bin's judgment is,Agricultural drone technology is in the middle stage of development,Agricultural drones and even the drone industry is not like the automotive industry"Volume",At present, in addition to price competition, there is room for innovation,The industry has not arrived"Color TV fight",Looking ahead,The dividends of drone technology are still two or three years。

The reporter learned that there are also different discussions in the industry about the maturity and technical space of agricultural drones. The above-mentioned engineer told reporters that he judged that the maturity of the UAV is not so high, and one level is one of the major parts of the user's cost, and the maintenance cost needs to be reduced. The maintenance cost is related to the service system and product performance, which is reflected in the maintenance rate, the maintenance rate of UAVs decreased by 40% last year, and the maintenance rate will be 20% or 30% in the future, and the maintenance time is basically within 6 hours, and it will decline in the future. In addition, from the perspective of cultivated land area, the current agricultural drones cover mainly large and medium-sized plots of cultivated land, and large and medium-sized plots with an area of more than dozens of acres account for about half of the plot area, which means that the other half of the small plots have not been covered by UAVs, in addition, the penetration rate of UAVs in cash crops is not high, and there are still many scenarios to enter.

In terms of function, agricultural drones are also considering the use of new technologies, and both DJI and XAG have used AI. However, the reporter learned that at present, AI technology is more used for obstacle recognition on agricultural drones, and has not been used to realize the fully automatic and intelligent operation of drones. An industry insider described to reporters that analogous to the classification of autonomous driving, AI is at most in the L2 stage on drones.

Peng Bin said that the company has trained a drone AI model, which can be used to identify pedestrians, obstacles and other specific features, but it does not use today's super-large models and super-large computing power, nor does it use end-to-end large models, and with the maturity of technology and cost reduction, it is expected that AI can be used on a large scale in drones in the future. It is expected that there will be one or two generations of products that can be optimized in terms of computing power, and then the experience upgrade brought by the increase in computing power will be weakened.

The above-mentioned engineer told reporters that DJI has plans to use AI and other technologies to achieve a greater degree of machine automation in the future. The reason why this year's products are equipped with AI functions for the first time is that there are hardware support factors behind it, and AI algorithms can only be applied after the capabilities of the lidar and computing power platforms meet the standards. From the perspective of technical space, unlike the relative plane of the car, UAV AI needs to consider an additional height dimension when avoiding obstacles, which is a big challenge, and it also needs to improve computing power and sensor accuracy. In the future, from L2 to L3, AI will not only learn to recognize obstacles, but also learn to automatically choose the optimal route to detour, changing the situation that drones stop when encountering obstacles and need to be taken over by humans.

"Today's drones are still semi-automated, and people still have to move with the equipment in the field, in fact, as the equipment becomes more and more efficient, the burden on people is getting heavier and heavier. The process of spreading fertilizer or spreading still accounts for at least 30% of the workload of the operation flow, and if you want to develop a fully automatic process that can replace manual dosing and battery replacement, the cost is still very high. In order to improve the performance of batteries and sensors in the future, a technological leap is needed to achieve the goal of unattended. Said the aforementioned engineer.

(CBN reporter Liu Xiaojie also contributed to this article.) ）

(This article is from Yicai)