Weifan Intelligence secures funding for brain-inspired robot chips

🔑 Enhanced Key Takeaways

•Weifan Intelligence was established in May 2025, originating from the Peking University Brain-like Chip Laboratory (PAICORE Lab).
•The company's strategic goal is to develop a fully domestic robot core computing solution, aiming to reduce China's current dependence on foreign-made chips, such as NVIDIA's Jetson series, for embodied intelligence applications.
•The seed funding round, amounting to hundreds of millions of yuan, was co-led by Zhongguancun Capital and its subsidiary Qihang Investment, with additional investment from Shanghai Future Industry Fund, Shixi Capital, BAW Storage, Yanchuang Group, Haiyi Investment, and Tanyuan Venture Capital.
•Co-founder Yin Jilei brings over two decades of experience in the semiconductor industry, having held significant roles in chip research and development at companies like IBM, GlobalFoundries, MTK, and VIA.
•The BiGPU architecture is specifically engineered to support state-of-the-art (SOTA) large models used in embodied intelligence, integrating both brain-like computing and general-purpose GPU capabilities.

🛠️ Technical Deep Dive

The BiGPU architecture integrates brain-like computing with general-purpose GPU computing capabilities, specifically designed for SOTA large models in embodied intelligence.
It achieves structural unification of Spiking Neural Networks (SNNs) and Artificial Neural Networks (ANNs) by sharing instruction formats and address addressing.
This unification allows for a single instruction set and software toolchain, which significantly reduces development complexity and ecological access costs compared to traditional heterogeneous systems.
Power consumption is reduced by incorporating a brain-like computing mechanism that converts standard neural networks to SNN-based accumulation.
Neuromorphic chips, in general, enhance efficiency by co-locating memory and computation, mimicking the synaptic organization of the mammalian brain to achieve low-latency and energy-efficient processing.
The conversion of ANNs to SNNs, a core aspect of this technology, aims to minimize conversion errors and maintain performance with low latency, although challenges include quantization and unevenness errors.

🔮 Future ImplicationsAI analysis grounded in cited sources

Weifan Intelligence's BiGPU could significantly accelerate the development and deployment of embodied AI in robotics.

By addressing the critical issues of high power consumption and cost in robot brains through a unified SNN/ANN architecture, it enables more efficient and practical integration of advanced AI into physical robots.

The company's success could reduce China's reliance on foreign suppliers for advanced robot brain chips.

As a 'fully domestic robot core computing solution,' Weifan Intelligence directly competes with foreign products like NVIDIA's Jetson series, aligning with national strategic goals for technological self-sufficiency.

The BiGPU architecture is likely to simplify the development process for embodied AI applications.

By offering a unified instruction set and software toolchain for both SNN and ANN, it lowers the complexity and cost associated with developing and deploying advanced AI models for robotics.

⏳ Timeline

2025-05

Weifan Intelligence founded, incubated from Peking University Brain-like Chip Laboratory (PAICORE Lab).

2026-05

Weifan Intelligence completes seed funding round of hundreds of millions of yuan.

Weifan Intelligence secures funding for brain-inspired robot chips

⚡ 30-Second TL;DR

Key Points

🧠 Deep Insight

🔑 Enhanced Key Takeaways

🛠️ Technical Deep Dive

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline

📎 Sources (7)

👉Related Updates