Alibaba Leads $300M Funding Round for AI Evaluation Startup UniPat

Alibaba Group is leading a $300 million funding round for UniPat AI, an AI training and benchmark testing startup, at a valuation of $250 million, according to a Bloomberg report on September 10. The round is expected to close soon, with existing investors including Tencent Holdings and Sequoia also participating, people familiar with the matter said. Negotiations are still ongoing and terms could change.

UniPat was founded by Li Kuan, who previously worked on post-training analysis, data synthesis, and reinforcement learning at Alibaba's Tongyi AI lab. The company designs testing and evaluation scenarios that simulate real-world usage for AI models. Established in late 2025, UniPat's benchmarks cover software engineering capabilities of AI coding agents, everyday web operation capabilities of browser agents, and visual reasoning capabilities of cutting-edge multimodal models.

The investment underscores the willingness of tech giants to pay premium prices for high-quality data and deeper visibility into AI system performance. With copyright and privacy restrictions tightening, the availability of human-generated internet data for large language models is shrinking rapidly. Developers also frequently encounter issues where model benchmark scores are inflated and do not match real-world performance. UniPat aims to address both the data and evaluation bottlenecks.