AI to ROI Big Story: China’s Open-Weight Model Companies Make Their Move

2026-01-20 · Source: AI to ROI - By Ray Rike and Peter Buchanan · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation, Cybersecurity & Data Privacy · Depth: Intermediate, medium

Summary

Chinese open-weight AI models have significantly surpassed US frontier models in token volume on OpenRouter, now accounting for 45% to 61% of top-model traffic, up from US models' 70% share a year ago. DeepSeek alone commands 16.3% of this volume. This shift is primarily due to superior price-performance, with DeepSeek V4 Pro priced as low as \$0.003625 per million tokens compared to GPT-5's \$2.50, and MiniMax's M3 matching GPT-5.5-class coding performance at 5-10% of the cost. Major US enterprises like Walmart, Uber, and Microsoft are implementing usage limits on US models, with Microsoft even considering DeepSeek on Azure. Concurrently, US government actions, such as temporarily shutting down Anthropic's Mythos and Fable, and IP concerns like Alibaba's alleged distillation campaign against Claude, have further driven interest in open-weight alternatives. Chinese labs are well-funded and technically advancing, while US labs respond with price cuts and trust initiatives amidst inconsistent government policy.

Key takeaway

For Directors of AI/ML or CTOs managing escalating AI costs and seeking resilient model access, you should actively evaluate Chinese open-weight models like DeepSeek, Z.ai, and Moonshot AI. These models offer significant price-performance advantages and deployability, enabling substantial cost reductions and greater control. Consider implementing a two-stack AI strategy, using premium closed models for regulated workloads and open-weight models for commodity tasks to optimize spending and mitigate geopolitical risks.

Key insights

Chinese open-weight AI models are rapidly gaining market share due to superior price-performance and deployability.

Principles

• Price-performance dictates enterprise AI adoption.
• Open-weight models enable self-hosting and modification.
• Geopolitical factors significantly shape AI market access.

In practice

• Evaluate Chinese open-weight models for cost savings.
• Explore self-hosting open-weight models for control.
• Diversify AI model providers to mitigate risks.

Topics

• Open-Weight AI
• Chinese AI Models
• AI Market Competition
• AI Cost Optimization
• AI Geopolitics
• DeepSeek

Best for: AI Engineer, Machine Learning Engineer, NLP Engineer, Director of AI/ML, VP of Engineering/Data, CTO

Related on AIssential

• Why US companies are ditching American AI models for Chinese ones - ThePrint — Chinese open-weight AI models are gaining traction among US companies due to lower costs and enhanced customization capabilities.
• AI Costs Are Surging and the Cheap Model Fix Might Not Last — Geopolitical shifts may end cheap open-weight AI models, forcing focus on efficiency and Western alternatives.
• China is falling behind in the AI race, according to a US government benchmark — Chinese AI models lag US counterparts in capability but offer significant price advantages.
• Kimi K3's Victory Over Anthropic Examined — Chinese AI labs are rapidly advancing large open-weight models, challenging established Western AI companies and regulatory frameworks.
• Open-Source AI Gains Ground as Rising Costs Push Shift to Smaller Models — Open-weight AI models offer substantial cost savings and near-parity performance compared to closed models, yet remain significantly underutilized.
• China Models Now More Than a Third of Enterprise Market — Chinese LLMs captured significant U.S. enterprise market share on OpenRouter, demonstrating price/performance's decisive role in commoditizing inference.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI to ROI - By Ray Rike and Peter Buchanan.