Goldman Sachs raises China AI revenue forecast 30% to $13b
Synopsis
Key Takeaways
Goldman Sachs has raised its annualised recurring revenue (ARR) forecast for China's artificial intelligence model market by 30 per cent to US$13 billion, citing aggressive price cuts, technical breakthroughs, and accelerating corporate adoption as of August 5, 2026. The upward revision — from a prior estimate of US$10 billion — reflects the rapid rise of domestic AI players, particularly DeepSeek and MiniMax, whose cost-efficiency gains are reshaping competitive dynamics across the sector.
Why It Matters
The revision signals a structural shift in how Wall Street is valuing China's AI ecosystem. In a research note published on Monday, August 4, 2026, the US investment bank highlighted that domestic AI pioneers have reached 'new frontiers in performance per dollar,' a metric increasingly central to enterprise procurement decisions globally. The scale of the upgrade — a full US$3 billion jump in projected run-rate revenue — underscores how quickly the market has evolved beyond initial expectations.
The Competitive Backdrop
Goldman Sachs lifted its projected year-end ARR for Hong Kong-listed Zhipu AI to US$2.5 billion and for MiniMax to US$1 billion. The bank noted that 'competition for the best performance-to-price balance' is set to intensify among Chinese AI model providers. These projections arrive as both companies demonstrate measurable traction in enterprise and developer markets.
The competitive pressure intensified markedly last week. On Friday, August 1, 2026, MiniMax launched its H3 model under an open-weight approach — a multi-modal system capable of processing text, image, video, audio, and music — priced at just 30 to 50 per cent of incumbent market levels, according to the bank's research. On the same day, DeepSeek officially opened API access for its V4 Flash model, targeting front-end coding capabilities that rival Zhipu AI's flagship GLM-5.2.
Market Reaction
According to Arena AI's latest global leaderboard, Zhipu AI's GLM-5.2 currently ranks seventh in front-end coding worldwide, with DeepSeek V4 Flash closely trailing in eighth place. The narrow gap between the two models illustrates how rapidly the performance differential between leading Chinese AI systems is compressing — a dynamic that could accelerate further price competition.
What's Next
The Goldman Sachs report frames the current period as one of intensifying price-performance rivalry rather than a winner-take-all consolidation. As open-weight models like MiniMax H3 lower barriers to enterprise adoption, incumbents including Alibaba, Tencent Holdings, and ByteDance face pressure to respond with competitive pricing or differentiated capabilities. Investors and enterprise buyers alike will be watching whether the US$13 billion ARR ceiling holds — or gets revised upward again before year-end.