The impact of "Radish Run" on the traditional economy is expected to be less than anticipated, with numerous connections to the world of cryptocurrency

1. The Impact of Robotaxi on Traditional Economy is Expected to be Less than Generally Anticipated

In early July 2024, after media reports on the operation of Robotaxi in Wuhan, it trended on Weibo. The structural impact of autonomous driving robotaxis, represented by companies like Baidu and Tesla, on the transportation industry has attracted attention from industry researchers and internet public opinion, with discussions extending to concerns about AI's potential to replace and displace labor employment positions in the future.

Since 2022, Robotaxi and Pony's autonomous driving taxis have begun small-scale operations in cities like Wuhan, Beijing, and Shanghai. As of April 19, 2024, Robotaxi has provided over 6 million rides on open roads, making it the largest autonomous driving mobility service provider in the world. In May of this year, Baidu launched the sixth generation of Robotaxi in Wuhan, equipped with the world's first autonomous driving large model supporting L4-level applications. According to Baidu, the Wuhan Pilot project is expected to achieve financial breakeven by the end of 2025.

Currently, the most media and internet attention in the domestic passenger charging test projects across 12 cities is focused on the trial deployment of Robotaxi in Wuhan. The scope of the trial operation in Wuhan has expanded significantly, covering most of the main urban areas and functional zones, as shown in the image below:

Currently, due to various factors such as subsidies and labor costs (no drivers are needed, only one remote safety officer is required for every three cars), economies of scale (uniform procurement of the same model, which reduces procurement and maintenance costs as the fleet size increases, and scaling laws can non-linearly enhance the computational efficiency of algorithm models with more data), and a robust supply chain (China's new energy and intelligent connected vehicle supply chain is quite strong, with a complete and high-level industrial chain for electric vehicles, charging, lidar sensors, and other components), some users report that the per-mile cost of Robotaxi is significantly lower than that of traditional ride-hailing and taxis. In addition to price, it also has the following advantages:

• Autonomous driving robotaxis avoid issues that may exist in traditional ride-hailing and cruising taxis, such as unpleasant odors, unfriendliness to women, and some drivers unwilling to use air conditioning to save costs, thus improving the riding experience. They also have a unified driving mode, avoiding issues like sudden braking and excessive honking by some drivers, enhancing safety, quietness, and comfort;

• All orders for autonomous driving robotaxis are automatically sorted by the system, avoiding chronic issues in the cruising taxi industry such as refusal of service, picking up passengers midway, intentionally taking longer routes, and inaccurate meter readings.

The advantages of the aforementioned robotaxi, combined with the strong policy support from the central to local governments for new energy, intelligent connected vehicles, and the autonomous driving industry, pose a certain challenge to the already saturated ride-hailing and traditional cruising taxi markets. However, due to various restrictions on the autonomous driving taxi projects of Robotaxi and other operators like Pony, such as access policies, operational regions, and specific pick-up and drop-off locations, it is currently difficult to see the possibility of a disruptive competition from robotaxi to the ride-hailing industry in the short term.

From a technical perspective, in the future, Baidu may follow the technological trends led by Tesla and enhance the performance of its autonomous driving large models by upgrading the current traditional modular stack models to globally trained end-to-end models, making the driving behavior of the large models more "human-like." However, end-to-end models face issues such as weak interpretability, difficulty in error attribution and debugging, and strong black-box characteristics. Given the strong regulatory conditions and low tolerance for failures in the autonomous driving industry in China, whether end-to-end models can be truly applied remains uncertain and may limit the further development of autonomous driving taxis.

The author believes that due to the many limitations of the robotaxi model represented by Robotaxi (especially the restrictions on pick-up and drop-off locations in urban areas), it is foreseeable that Robotaxi will not be able to truly replace human-driven ride-hailing and cruising taxis in the near future.

Currently, the number of vehicles deployed by Robotaxi in Wuhan is less than 500, which is still far from posing any substantial threat to the traditional taxi model, which has over 16,000 traditional cruising taxis in Wuhan. It seems that the protests from the taxi industry in Wuhan may be just a habitual reaction similar to their excessive response to shared bicycles. From the author's personal experiences in cities like Beijing and Wuhan, the poor service attitude and driving habits of traditional taxi services are quite severe, and there has not been much improvement even after the introduction of ride-hailing services. Perhaps the external impact from the L4 autonomous driving taxi model represented by Robotaxi can promote improvements in the traditional taxi model and drive structural enhancements in local tourism, consumption, and other tertiary industries. Media and public opinion are more focused on the potential replacement of human drivers by AI autonomous driving but overlook the role of the robotaxi model in optimizing the service industry economic structure and alleviating friction in transportation infrastructure within the economic operation, possibly overestimating the impact of robotaxi on the macroeconomy. The author believes that areas like Xiong'an New Area and Beijing Yizhuang, where traditional operating vehicles are lacking, and the urban areas are relatively new with simple road conditions, are more likely to be regions where the robotaxi model can expand its business.

2. L4 and Above Autonomous Driving is Here, Strengthening and Expanding the DePin Data Infrastructure Track for Cryptocurrency

The recent surge in popularity of Robotaxi in China has unveiled the commercial application of L4 and above level autonomous driving, while the data concept within the core three elements (computing power, algorithms, data) of AI large models supporting autonomous driving is closely linked to the cryptocurrency industry. Distributed computing resources and data have become one of the most promising concepts and tracks in the cryptocurrency industry since 2023, and blockchain distributed technology and DePin projects provide possible solutions to many issues related to AI computing power and algorithms. For instance, Hivemapper addresses the high costs and difficulties in obtaining training data for models: whether using traditional modular vehicle-cloud stacks or the latest end-to-end technology, LLMs require massive labeled data for training to meet autonomous driving needs, especially with higher requirements for video labeled data, which typically has a higher unit price among various data resources.

Hivemapper provides data services that include video and map information to AI companies. As a decentralized global mapping initiative, Hivemapper leverages blockchain and community participation to create a detailed, real-time, and easily accessible mapping platform. Users record map information through onboard cameras (dashcams) and upload it to Hivemapper's open-source network, receiving HONEY tokens as rewards in return. To enhance network efficiency and reduce interaction costs, Hivemapper has chosen to build its platform on the Solana blockchain, anticipating better integration with autonomous driving and large model narratives in the future.

Reference

1. https://www.theblockbeats.info/news/49070

2. https://research.mintventures.fund/2024/04/08/zh-a-new-solana-based-ai-depin-project-a-brief-analysis-of-upcoming-tokenlaunch-io-net/

3. https://mp.weixin.qq.com/s/qyA6W_Jplo-Sb89kGglqqg

END

This content is for informational sharing only and does not promote or endorse any business or investment activities. Readers are advised to strictly comply with the laws and regulations of their respective regions and not to engage in any illegal financial activities. No trading entry, guidance, or issuance channel for any virtual currency or digital collectibles is provided.

4Alpha Research content is prohibited from being reproduced or copied without permission. Violators will be held legally accountable.