China HD Map Evolvement — Ultimate Data Ownership
HD map is an exciting space that emerges with autonomous vehicles technologies as it touches upon every component in autonomous driving, and combines engineering efforts from surveying & mapping, autonomous driving system and the cloud. Moreover, to survive a cash-burning and time-consuming space with intense competition eventually comes down to a smart and sustainable business model.
Map history and 3 value chains
In China, a “Grade A Surveying and Mapping Certificate” is required to legally collect geographical data and create digital maps for commercial use. Since 2001 when NavInfo obtained the very first certificate, 22 companies got certified in total.
Over the last 20 years, three value chains have been presented at different times. In the early days, automobile navigation was the goal and it was digital map companies who got the certificate. They created maps with expensive equipment and profited from licensing as well as updating services and location based services.
From 2007 to 2017, the space for a mapping certificate was quiet with only 2 public institutions and 1 independent company. This was also the period of AV technologies progressing rapidly, soon after which the demand for HD map emerged. Between 2017 and 2019, the certified players mainly aimed for AV road test and future commercialization, including HD mapping startups, ride-hailing platform, OEM, and Tier 1 (Huawei).
Starting late 2019, the value switched to autonomous delivery (robodelivery). SF express and JD have both obtained the certificate and Meituan’s application is in process.
List of companies with “Grade A Surveying and Mapping Certificate”
Robodelivery proposition in HD map
Various companies have taken on the HD mapping business with the most recent as delivery.
Robodelivery is about carrying goods on a relatively fixed route at low speed making its commercialization easier and faster than robocars. Meanwhile, the COVID-19 outbreak gave last-mile delivery a scene to showcase its additional value. JD, SF express, Meituan, Suning, etc. have all participated and determined on a continued investment even after the virus spread.
As robodelivery penetrates China, these companies will be increasingly reliant on customized HD map for the last mile areas. One could argue that with all the operational data collected, map vendors could serve the needs on top of the standard solutions. However, as data means much more, another option for delivery players is to tackle mapping on their own.
For example, SF Express’ map product has integrated with the entire delivery process, managing 700 million pieces of location tags and nearly 100 million km of trajectory data on a daily basis. These data could be transformed into commercial products serving the government, applicable in oil, electricity, public safety emergency system, natural resources and city management, etc.
The fact that major delivery companies are now able to conduct legitimate business around map data in a closed loop is challenging the status of traditional map vendors.
Ultimate winners
Robodelivery companies have rolled out their plans but the same solution certainly does not apply to robocars. The puzzle for AV is still circulating among map vendors, OEMs, Tier 1s, and startups, which have developed different approaches to HD mapping over the recent years.
Map vendors including tech giants’ subsidiaries usually go for the traditional way by using single cost of million-dollar equipment for data collection to produce a base map. Startups tend to leverage affordable hardware and software solutions like computer vision for comparable precision. For startups without the certificate, they offer solutions and services to transform fleet data into HD maps.
For map updates, crowdsourcing is a commonly recognized solution but there are practical issues to be solved. Though this is not an urgent matter given the vehicles are still L2 to L3 automated with humans involved for a very long time.
Despite different approaches, HD mapping is still a highly complicated, laborious, and expensive process. Here is the P&L issue. AutoNavi, one of the map vendors also as subsidiary of Alibaba, announced the plan to charge no more than RMB 100 (USD 15) per car per year for OEMs to plug their map to mass-produced vehicles. This is more like an aggressive strategy to gain market share, but profits still have to be made.
Just as in robodelivery, data is probably the end game, which can be monetized and turned into bulk profits. The ones with access to the most data would own the most power and potentials.
Next is the key debating area — map vendors or OEMs? Technically, it is the OEMs who own most of the data, because they constantly collect with their onboard sensors. They are unlikely to give it up to third parties.
The future of China HD map landscape is not so clear yet. But we know the criteria of ultimate winners: technologies that can scale in real world and a smart business model around data ownership.
