ByteDance's New AI Data Department: A Strategic Move to Strengthen Core Model Training
ByteDance is doubling down on its AI ambitions with a significant organizational shake-up. The tech giant has reportedly established a new first-level department named "AI Data and Security," placing it on equal footing with existing units like Seed, Flow, and Douyin. This new entity is being led by Wang Yinglei (also known as Adam Wang), who previously headed the TikTok platform responsibility team and oversaw TikTok live streaming. As of this writing, ByteDance has not issued an official response to these reports.
According to insiders, the department's roots trace back to a global data team founded in 2023 by Fuyue (Yuyi), a member of TikTok's founding team. That team, initially around 100 people strong, has now been merged with other data-focused groups, including the group's data middle platform DMC and the AI data platform AIDP under Flow. These previously scattered units are being consolidated under one roof, a process that began in early June 2026 and is still ongoing as the company fine-tunes its organizational structure and personnel.
The newly formed AI Data and Security department is poised to become a sprawling data organization that cuts across both base model development and business teams. Its core mission? To provide comprehensive, cross-modal data services for all of ByteDance's large models. This covers the entire data lifecycle—from setting standards and sourcing data to synthesizing, cleaning, and evaluating quality. In essence, it's about ensuring that every model gets the high-quality data it needs to perform at its best.

This restructuring comes at a time when ByteDance is pouring ever more resources into artificial intelligence. Earlier reports from 36Kr indicated that in 2026, the company plans to focus on world models, video generation, coding, and the commercialization of its Doubao product, with corresponding increases in investment in training data. The creation of this new department underscores a strategic shift: data is no longer just a supporting element but a critical piece of infrastructure in the race to build superior AI models.
By consolidating data resources and production capabilities, ByteDance is positioning itself to handle the demanding training needs of multi-modal models and diverse business applications. It also signals a growing recognition that data engineering and data quality are becoming decisive factors in the competitive landscape of large models. As the AI arms race heats up, having a dedicated, well-organized data powerhouse could give ByteDance a significant edge.
For now, the industry will be watching closely to see how this new department evolves and what impact it has on ByteDance's AI roadmap. One thing is clear: the company is betting big on data as the foundation for its future AI breakthroughs.
Key Points
- ByteDance has established a new first-level department, "AI Data and Security," led by Wang Yinglei.
- The department merges several existing data teams, including the global data team, DMC, and AIDP.
- Its primary role is to provide cross-modal data services for all ByteDance large models.
- The move reflects ByteDance's increased investment in AI and the growing importance of data quality.