Zhipu earned $110M but spent twice that on R&D — can open-weight models pay the bills?
ccerrato147 · x · 2026-11-09
A thread by @ccerrato147 breaks down Zhipu's (Z.ai) business model:
- GLM is a family of open-weight models anyone can download and run locally; the company earned $110M last year while spending more than twice that on R&D
- Zhipu listed in Hong Kong in January at $7B; within five months its market cap hit $170B — nearly half of Alibaba — then shed about $130B by late September
- It began as a research tool on a single desktop at Tsinghua; the core tension is giving its best models away for free
- Two monetization paths: cloud API access and on-prem installs — free downloads still need GPUs, uptime, security review and 3am support. "That gap is the business. So far it does not close."
- The bet: open-model reputation converts into contracts nobody else can serve; the question is how long a company can spend 2x its revenue making it
Source: Asianometry video, 2026-11-01, 22 min.
Related event: Anatomy of Zhipu's Open-Source Gamble: $110M Revenue, Volatile Valuation(13 posts)→
More from Models
- Model recipe via rigorous experiments, but scaling could rewrite the story — liliyu_lili · 2026-11-09
- Artificial Analysis launches Decision Index v2 with 45% private data and calibration-based scoring — ArtificialAnlys · 2026-11-09
- GLM-4.5 Hit #1 on Hugging Face Within 48 Hours of MIT-Licensed Release — ccerrato147 · 2026-11-09
- Has any LLM release ever been decidedly worse than its predecessor? — adonis_singh · 2026-11-09
- CEOs say Google's Argon stands out for writing, calling Google the AI dark horse — danshipper · 2026-11-09
- Epoch AI adds ECI scores, projects Gemini 3 Pro time horizon at 4.9 hours — CFGeek · 2026-11-09