Qiming Venture Partners Portfolio Company Tashi Zhihang Unveils World's First Scalable Real-World Embodied Multimodal Dataset WIYH

Qiming Venture Partners portfolio company Tashi Zhihang has unveiled World In Your Hands (hereinafter referred to as the "Tashi WIYH Dataset") — the world's first large-scale real-world embodied VLTA (Vision-Language-Tactile-Action) multimodal dataset — and plans to open it to the industry in December 2025. This milestone marks the formal establishment of Tashi Zhihang's industry-first Human-Centric embodied data engine paradigm, a technical approach that is roughly six months ahead of Tesla's Optimus.

For a long time, the internet data and simulation data that mainstream large models rely on for pre-training have suffered from two shortcomings: internet data varies wildly in quality and lacks action information; simulation data has limited authenticity and poor scene generalization, making it difficult for trained models to smoothly transfer to the real world. For humanoid robots, the biggest hurdle on the path to "embodied intelligence" is not the algorithm itself, but how to obtain scalable, authentic, and generalizable data. The scarcity of high-quality training data is a widely acknowledged challenge in the industry.

"The release of the Tashi WIYH Dataset marks the first time the industry has collected vision, language, tactile, and action multimodal data at scale across real-world industries and tasks," said Wenchao Ding, Chief Scientist at Tashi Zhihang. "It lays the foundation for achieving the scaling law of embodied foundation models in the future."

In the Human-Centric first-person data collection videos released by Tashi Zhihang, the difference is clear: unlike the static, monotonous environments of previous labs and data collection factories, the Tashi WIYH Dataset draws on multiple real-world industry work settings and personnel, capturing standard operating procedure data across embodied scenarios including hotel laundry, supermarket restocking, and logistics operations. The Tashi WIYH Dataset doesn't just solve the problems of "insufficient data volume, low quality, and high costs" — it ensures the data truly "comes from the real world."

A laundry worker wearing self-developed collection equipment folds towels

The Tashi WIYH Dataset has four defining characteristics:

• Authentic: Collected from real embodied tasks, closely aligned with actual model application scenarios;

• Rich: Spanning multiple industries and manipulation skills, giving models transfer and generalization capabilities and breaking down data reuse barriers;

• Comprehensive: Encompassing full ground-truth multimodal data across vision, language, tactile, and action, facilitating pre-training modality alignment;

• Massive: Scale comparable to large language models, ensuring future potential for embodied intelligence.

Based on these four core characteristics, the Tashi WIYH Dataset delivers three distinctive advantages:

• First, in modality completeness: Through self-developed collection hardware, it synchronously captures vision (RGB), force-tactile (pressure sensor signals), and action (finger joint poses and end-effector trajectories) data, ensuring precise temporal and spatial alignment of multi-source data;

• Second, in data annotation pipeline: The Tashi WIYH Dataset leverages proprietary cloud-based foundation models for high-precision annotation, covering multi-granularity ground-truth labels including 2D semantics, scene depth, operational task decomposition, affordance of interactive objects, and hand and end-effector motion trajectories, providing comprehensive and multi-dimensional supervisory signals for embodied foundation model pre-training;

• Finally, in collection environment: Tashi Zhihang goes deep into real-life operational scenarios. Compared with the industry's typical high-cost proprietary data collection and training factories, it collects standard operating procedure data from workers in non-constructed, non-dedicated, non-enclosed environments — significantly improving data authenticity, diversity, and generalization while reducing collection costs by more than an order of magnitude.

A supermarket worker wearing self-developed collection equipment restocks shelves

The introduction of the Tashi WIYH Dataset marks the establishment of a Human-Centric embodied data paradigm. It makes World Engine pre-training for real-world embodied AI possible. Rooted in "thousands of industries," the Tashi WIYH Dataset is expected to achieve "one model, thousand tasks," becoming critical corpus and infrastructure for training general embodied foundation models, pushing industry applications from single-task execution toward general manipulation capabilities, and laying a solid foundation for embodied robots to truly enter thousands of enterprises and households.

---

Past Coverage

Qiming Stars | Tashi Zhihang Raises $120 Million in Angel Round, Qiming Venture Partners Co-Leads Dialogue | Alex Zhou of Qiming Venture Partners in Conversation with Yilun Chen of Tashi Zhihang and Wenbin Tang of Yuanli Lingji: Macro Consensus and Micro Non-Consensus on Embodied Intelligence Qiming Honors | Qiming Venture Partners Selected for Zuikeji T50 — China's Top 50 Pioneer Tech Investment Institutions

Founded in 2006, Qiming Venture Partners currently manages 11 USD funds and 7 RMB funds, with total assets under management reaching $9.5 billion. Since its inception, the firm has focused on investing in outstanding early- and growth-stage companies in Technology and Healthcare.

To date, Qiming Venture Partners has invested in over 580 high-growth innovative companies, of which more than 210 have gone public on the New York Stock Exchange, NASDAQ, Hong Kong Exchanges and Clearing Limited, Shanghai Stock Exchange, and Shenzhen Stock Exchange, or exited through M&A and other means. Over 80 portfolio companies have become recognized unicorns or super-unicorns.

Many Qiming Venture Partners portfolio companies have grown into the most influential players in their respective fields, including Xiaomi (01810.HK), Meituan (03690.HK), Bilibili (NASDAQ:BILI, 09626.HK), Zhihu (NYSE:ZH, 02390.HK), Roborock (688169.SH), UBTECH (09880.HK), WeRide (NASDAQ:WRD), Insta360 (688775.SH), Unisound (09678.HK), Gan & Lee Pharmaceuticals (603087.SH), Tigermed (300347.SZ, 03347.HK), Zai Lab (NASDAQ:ZLAB, 09688.HK), CanSino Biologics (688185.SH, 06185.HK), Schrödinger (NASDAQ:SDGR), MicroPort EP MedTech (688617.SH), Sanyou Medical (688085.SH), Amoy Diagnostics (300685.SZ), Berry Genomics (000710.SZ), Sinocelltech (688520.SH), Yuanxin Technology, Insilico Medicine, MediLink Therapeutics, LaNova Medicines, Zhipu AI, StepFun, Biren Technology, and others.