The collaboration between DiDi and NVIDIA brings China's ride-hailing giant one step closer to realizing a safe and grand-scale application of autonomous driving technology.
Ride-hailing giant DiDi Chuxing announced on December 18 its collaboration with NVIDIA to further develop its autonomous driving technology and cloud computing solutions.
DiDi has said that the company will be using NVIDIA’s GPUs in its data center to enhance machine learning algorithms. Another technology that will be utilized is NVIDIA DRIVE, which will increase the inference capabilities of its L4 autonomous vehicles.
The advanced technology that NVIDIA built is capable of integrating data from multiple sensors by adopting deep neural networks, to interpret a 360-degree surrounding environment, hence delivering the safest driving route.
In terms of cloud services, DiDi will be supported by vGPU's license model to construct leading AI infrastructure and vGPU cloud servers that provide better user experience and broader application scenarios.
DiDi has enjoyed an outstanding growth rate over the past few years. The company acquired Uber China back in 2016, and is currently handling a staggering 550 million users on its platform. It has penetrated global markets such as Australia and Mexico, and is eyeing on Costa Rica as its next expansion plan.
Despite this, DiDi is sharing the same fate as its global rivals Uber and Lyft, as it struggles to generate profits, recording a loss of USD 1.6 billion in 2018. It was estimated that as much as USD 1.67 billion was poured into driver subsidies. The company has not been able to sustain the cash-burning model that it adopted for its rapid expansion.
New contenders have also begun emerging in the ride-hailing market with its more attractive commission rates. These companies include Meituan-Dianping and the state-owned BAIC Group (Beijing Automotive Industry Co).
In addition to this, DiDi was caught up in several scandals, such as the rape and murder cases suffered by female passengers when using DiDi’s hitching service.
DiDi has yet to catch up with its domestic competitors in terms of automated drivers. The Softbank-backed company is not only is relatively late in securing a self-driving test license in California but also lags behind in testing compared to its rivals. According to a report published in 2018, total automated driving test mileage in Beijing reached 163,600 kilometers, with Baidu as the leader, accounting for 91% of all mileage covered in China. This completely undermines the achievement of DiDi, which registered only 78 kilometers with its two vehicles.
DiDi has made several adjustments with regards to its autonomous driving unit over the past few months. Following the steps of its rival, Uber, DiDi upgraded its autonomous driving business earlier last August into an independent department, to focus on autonomous driving research and development. Since then, it has also begun cooperating extensively with industries in its value chain.
With DiDi removing over 300,000 unqualified drivers, it has managed to achieve only 65% of China’s request. Previously, DiDi had also successfully launched robo-taxis in Shanghai, which is a driverless ride-hailing service, aiming to tackle the driver shortage issue and close the supply and demand gap.