Data Lake / Data Warehouse Modernization
NIO's autonomous-driving program was generating tens of petabytes of data, and growing vehicle-data ingestion plus training-platform access frequency was making public-cloud network and storage costs, efficiency, and compliance constraints increasingly hard to manage.
Reported impact
NIO met autonomous-driving capacity, performance, compliance, and cost targets with a private-cloud storage platform that supports rapid data growth, EB-scale planning, faster AI-data access, and materially lower hardware spending.
NIO used Tencent Cloud's distributed storage stack to build a private-cloud foundation for autonomous-driving data instead of keeping all data on public-cloud storage. TStor OneCOS massive object storage is the core layer, giving NIO a data-lake and data-center storage base that can scale to trillions of objects and rapid capacity growth. Tencent paired OneCOS with GooseFS so data preparation, model training, inference, and small-file or HDFS-style access could run faster against the same shared pool. YottaStore provides the distributed storage engine with EC 8+4 redundancy, log cache, automatic balancing, and repair, while StarLake servers provide the dense hardware used to scale the platform into the hundreds of petabytes. The result is a hybrid/private storage environment for sensitive data, AI workloads, and lower-cost expansion.
Vehicle data and training-platform datasets land in TStor OneCOS, where YottaStore manages durability and cluster balancing. GooseFS then accelerates small-file and HDFS-style access so data preparation, model training, and inference can work against the shared storage pool, while StarLake servers provide the dense hardware layer that scales the environment to hundreds of petabytes.
Because sensitive data also needed to stay in a local data center for compliance, NIO chose to build its own private-cloud storage (original: Also due to compliance requirements, sensitive data also needs to be stored in the local data center. Therefore, NIO choose to build their own private cloud storage).
NIO had been using public-cloud storage for tens of petabytes, but rising vehicle-data and training-platform access frequency was making network and storage efficiency and cost more significant (original: NIO have been using public cloud storage for tens of petabytes of storage… the data collected by vehicles and the data of the training platform need to access the public cloud increasing frequency, and the impact of network and storage on efficiency and cost is becoming more and more significant).
TStor OneCOS supports unlimited expansion of a single cluster and trillions of objects in a single bucket, and with GooseFS it supports the full data-service process for data preparation, model training, and reasoning (original: Tstor OneCOS massive object storage adopts… supports unlimited expansion of a single cluster and trillions of objects in a single bucket… combined with big data solutions such as GooseFS, it supports data preparation, model training and reasoning full process data service).
The platform supports 1 trillion+ file indexes, 100GB+ bandwidth, 1 million+ QPS, 12 nines reliability, and average site expansion of 20GB/s (original: realize 1 trillion+ file indexes, support 100GB+ read and write bandwidth and 1 million+ request QPS… ultra-high reliability of 12 nines… the average speed of customer site scale expansion reaches 20GB/s).
Total hardware-resource cost was reduced by more than 50% (original: 总成本方面，硬件资源综合成本降低50%以上).
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Baidu AI Cloud Sample May 16, 2025 Geely Holding Group (吉利控股集团) Geely needed a group-wide cloud and data foundation that could support global manufacturing, R&D, and vehicle-intelligence initiatives while protecting internal data, improving collaboration, and giving business teams a… 1,000+ component suppliers collaborate on manufacturing 1+6+N Geely Hybrid Cloud Platform · Baidu Private Cloud · Cockpit Large Model 
Microsoft Sample Nov 11, 2024 JATO Dynamics To further support customers, JATO Dynamics wanted to create a reporting tool to offer car dealerships insights into their own performance compared to the market, including metrics like sales price, discount rates, and… By automating content generation, Crow says, dealerships are saving approximately 32 hours each month, allowing… Azure AI Search · Azure App Service · Azure Data Factory 
Tencent Cloud Sample China Media Group (Yangshipin) Yangshipin needed a secure, data-driven media architecture that could unify user, content, and production data while supporting large-scale video distribution, AI-assisted production, and 5G-era media operations. The dedicated cloud environment uses two availability zones for disaster recovery. AI Middle Platform · Big Data Platform · Data Middle Platform 