☀️ AI Morning Minute: Zhipu
The Chinese AI lab that became the first publicly traded frontier model company
Most of the big AI labs making headlines (OpenAI, Anthropic, Google DeepMind) are still private companies funded by venture capital or tech giants. Zhipu took a different route. They went public on the Hong Kong Stock Exchange in January 2026, beating every Western lab to the market and making a statement about how Chinese AI companies plan to compete.
What it means:
Zhipu AI is a Chinese AI lab founded in 2019 as a spinoff from Tsinghua University, one of the country's top research schools. The company builds the GLM series of large language models (GLM stands for General Language Model) and has rebranded internationally as Z.ai. In January 2026, Zhipu held its IPO on the Hong Kong Stock Exchange, becoming the first major frontier AI lab to go public. The offering raised about $558 million at a $7.1 billion valuation.
Why it matters:
Zhipu is proving Chinese AI can run on Chinese hardware. Their recent models were trained on Huawei Ascend chips using Huawei’s MindSpore framework, with no Nvidia GPUs involved. That matters because US export restrictions have made Nvidia hardware hard to get in China, and Zhipu’s success shows the domestic chip stack can compete at the frontier.
Their models are open source under the MIT license, which means any company or developer can download them, run them, and build on top of them for free. That’s a sharp contrast to OpenAI and Anthropic, whose best models are locked behind paid APIs. Zhipu is betting that giving the model away builds a bigger ecosystem than charging for it.
Going public changes the incentives. As the world’s first publicly traded foundation model company, Zhipu now has to answer to shareholders and publish earnings. That creates pressure to monetize, which is why they’ve raised prices twice in 2026 already. It also creates a governance model that OpenAI and Anthropic haven’t faced yet.
Simple example:
You're watching three kids build rival lemonade stands. Two of them (the American ones) are funded by rich parents who keep writing checks and telling them not to worry about profits yet. The third kid (Zhipu) sets up a booth, takes money from every neighbor who buys a glass, opens the books so everyone can see the recipe, and uses the cash to buy better lemons.
Different strategy, same goal: build the best stand on the block. The question is which approach wins when the neighborhood has to pick a favorite.