Alibaba Backs Former Employee’s AI Testing Startup at $2.5B Valuation
Synopsis
Alibaba is reportedly leading a $300 million investment in UniPat, valuing the AI testing startup at $2.5 billion as demand grows for reliable training data, realistic benchmarks and better ways to measure AI models’ real-world performance.
Alibaba is reportedly leading a $300 million late-stage investment in UniPat, an AI training and benchmarking startup founded by a former Alibaba researcher, in a deal that could value the company at about $2.5 billion.
Tencent and existing investor HSG, formerly known as Sequoia China, are also participating in the round, according to people familiar with the matter.
The talks are still ongoing, meaning the investment size and terms could change before completion.
UniPat is building infrastructure around one of the biggest challenges faced by AI developers today which is how to generate reliable data and accurately measure what increasingly capable models can actually do.
Building Better Tests for AI Models
UniPat creates realistic environments and scenarios that allow AI systems to be tested across different tasks. Its work includes benchmarks for coding agents, browser-based agents and multimodal models that need to understand visual information.
The company then sells the resulting training and benchmarking data to AI researchers and businesses. This puts UniPat in a growing part of the AI ecosystem that sits between model development and real-world performance.
Founder Li Kuan previously worked at Alibaba’s Tongyi AI lab, where his work included post-training analysis, data synthesis and reinforcement learning. UniPat was established in late 2025 and has also developed smaller models for prediction and scientific research.
Why AI Data Is Becoming More Valuable
The investment reflects a broader shift in the AI industry. Developers are facing growing limits on the amount of high-quality human-generated data available online, while copyright and privacy concerns are making some traditional sources harder to use.
There is also growing concern that public AI leaderboards do not always reflect how models perform in practical situations. Realistic testing can therefore become valuable in its own right.
The reported $300 million round would rank among the largest late-stage venture investments globally. It also highlights how Chinese technology giants are increasingly backing companies that can strengthen the data, evaluation and infrastructure layers around AI.
Alibaba has been expanding its own AI ambitions, including its focus on AI agents and model-as-a-service products. The company and Tencent have also repeatedly invested in China's emerging AI ecosystem.
For UniPat, the reported valuation would place it among a new generation of AI infrastructure companies whose value comes not from building a single chatbot, but from helping other AI systems become better, more measurable and more useful.
Source: Dealroom
Vishal is an experienced Editor at Inspirepreneur Magazine with key interests in artificial intelligence, eCommerce, entrepreneurship, lifestyle and startup sector. Prior to joining Inspirepreneur, he was a Content Writer cum Correspondent at Siliconindia Magazine, where he worked on Company Profiles, Cover Stories, Executive Profiles, Feature Articles and Thought Leadership content.
You Might Also Like
Elastic agrees to buy AI testing startup DeductiveAI in deal worth up to $85M
Oil prices hit highest since 2022 as Iran war disrupts global supply
