TL;DR
Z.ai (formerly Zhipu) is on track to become the first independent Chinese AI firm with about $1bn in annual sales, per Bloomberg. The angle is commercialisation: Z.ai is monetising at scale while open-sourcing its best models, an approach the cash-burning Western labs have struggled to match. The piece flags that the $1bn figure is a forward projection leaning partly on annualised recurring revenue, that the company is still lossmaking, and that much of its revenue comes from state-owned buyers.
One of China’s leading AI startups is closing in on a milestone the rest of the industry keeps missing. Z.ai, the company behind the GLM models, is set to become the first independent Chinese AI firm with $1bn in annual sales, Bloomberg reports.
The claim is a projection, not a booked result. But the trajectory behind it is real, and it points at something the West has struggled to copy.
Z.ai is making money at scale while open-sourcing its strongest models. That combination is not supposed to work.
The numbers
The base is small but the growth is not. Z.ai’s 2025 revenue was about 724m yuan, roughly $100m, up 132% on the year.
The forecasts are steep. JPMorgan expects 2026 revenue of about 4.6bn yuan, rising to 30.9bn yuan by 2028, the year it projects the company will finally turn a profit.
The billion-dollar line needs one caveat. It leans partly on annualised recurring revenue, a run-rate snapshot, rather than a full year of booked sales, and even JPMorgan’s 2026 revenue forecast sits below it in dollar terms.
How it actually makes money
The mix is enterprise-heavy. A large share comes from on-premises deployments for state-owned enterprises and financial institutions, alongside a fast-growing cloud business.
The API side is the momentum story. Annualised recurring revenue from its open platform reached 1.7bn yuan, up sixtyfold in a single year.
The open-source paradox
Here is the part that confounds the Western playbook. Z.ai releases its most capable models, including GLM-5.2, as open-source software anyone can download and run for free.
Giving the model away is supposed to destroy the ability to charge for it. Z.ai is betting the opposite, that free models drive adoption, and adoption sells cloud, support, and on-premises deployments.
It is the instinct its founder Tang Jie has defended in public, arguing frontier AI should stay open to everyone. The revenue figures are the commercial case for that philosophy.
Why it matters
Most AI companies, American and Chinese, lose enormous sums. Z.ai is no exception yet, and its losses have kept climbing even as revenue soars.
But approaching $1bn in sales is a different order of maturity from the pure cash-burn most labs live in. It suggests China’s machine for turning technology into revenue is now firmly pointed at AI.
TNW has argued that China’s real edge is commercialisation, not subsidies, the knack of scaling and monetising faster than anyone. Z.ai is that thesis applied to large language models.
The caveats worth keeping
The billion-dollar number is a forward projection, and projections in AI have a short shelf life. The company is still lossmaking, and much of its revenue leans on state-owned buyers, which blurs the line between commercial demand and state support.
The valuation is extraordinary too, at roughly $112bn after a rally of well over 1,000% since its January listing. Z.ai has already raised billions in a follow-on share sale, and the price assumes the projections come true rather than reflecting where the business is now.
It also operates in a brutal market, with cheap Chinese models undercutting each other and the US labs on price. Reaching $1bn in sales is one thing, and keeping enough margin to profit from it is another.
Still, the milestone would be a real one. Chinese AI’s story so far has been capability catching up and capital pouring in, and revenue at this scale is the first sign it can also start paying for itself.