Zhipu said on a call with analysts and investors that it has secured fresh financing to expand computing capacity after a surge in demand for its GLM-5 model strained internal resources and led the company to suspend its Coding plan.

According to the company’s remarks, demand for model calls rose tenfold after the release of GLM-5 in February. Zhipu said that jump rapidly depleted its computing power reserves, forcing it to suspend its main Coding plan product while it worked to expand capacity.

Zhipu said it completed a refinancing round in July and obtained another financing on September 11 to continue building out its compute footprint. The company presented the additional capital as support for both rapid business growth and ongoing development of larger-scale models.

Different sources describe the matter differently, and the relevant details still require official confirmation.

Separate reporting and company-related records support that Zhipu completed an H-share placement in July 2026 raising about HK$31.411 billion, with proceeds earmarked for areas including research and development, computing capacity, cloud services, business expansion, strategic investment and general corporate purposes. That supports the broader direction of the company’s statement, but does not fully confirm the later financing figure cited in the call.

Why It Matters

Zhipu’s update underscores a central feature of the current AI race: model demand can scale faster than infrastructure, forcing companies to raise capital quickly just to maintain service capacity and product rollout. That makes compute access and financing ability core competitive factors alongside model quality.

For the wider market, the episode is another sign that AI development is becoming more capital-intensive, with funding increasingly tied to cloud capacity, technical services and commercialization. Even outside crypto, that dynamic matters for AI-linked digital asset narratives built around compute, inference demand and infrastructure scarcity.