Goldman Sachs Highlights Tencent’s Hunyuan Hy4 as a Major AI Leap
TREE NEWS reports: In a research report published on August 31, Goldman Sachs analysts led by Ronald Keung and Lincoln Kong highlighted the release of Tencent’s Hunyuan Hy4 preview model, calling it the largest generational capability jump in the model family’s history. The report, covered by 华尔街见闻, emphasizes that Hy4’s early release—arriving about two months after Hy3—underscores Tencent’s rapid iteration pace. Notably, the model’s coding capability surged to 6th place on the Code Arena WebDev global leaderboard, up from 28th for Hy3, positioning Hunyuan back among the top-tier open-source models.
Market Impact Analysis
Tencent Stock (0700.HK): Goldman Sachs maintains a 12-month SOTP-based target price of HKD 670, implying roughly 47.2% upside from the current price of HKD 455.20. The analysts view Hy4’s progress as a key driver for Tencent’s stock over the next year, alongside strong WorkBuddy user metrics and the gradual rollout of WeChat Xiaowei. However, near-term earnings pressure is expected due to rising capital expenditure and AI investments, with EPS growth forecast at just 4% and 0% year-over-year for Q3 and Q4 2026, respectively.
Global Tech Sector: The news reinforces the intensifying competition in AI model development, particularly among Chinese tech giants. Tencent’s focus on a ‘product-model closed loop’—distributing models through its own ecosystem (WorkBuddy, CodeBuddy) and using real user feedback to iterate—could set a new standard for AI differentiation, potentially influencing how investors value AI-driven platforms globally.
AI Infrastructure and Costs: Despite the 2.6x increase in model parameters (from 295B/21B active to 770B/49B active) and expansion to 1 million token context length, Hy4 maintains competitive cost efficiency at roughly $0.45 per million tokens. The architectural innovations (Gated DSA Attention, IndexCache, iHC) suggest that Tencent is prioritizing inference efficiency, which could pressure competitors to optimize costs as well.
Key Takeaways for Investors
- Watch Tencent’s AI monetization: The closed-loop strategy could accelerate AI adoption in productivity and coding, potentially driving long-term revenue growth beyond near-term capex drags.
- Monitor upcoming events: Tencent’s participation in Goldman Sachs’ Asia Leaders Conference on September 1 will be closely watched for updates on compute allocation, depreciation pressure, and AI-driven advertising.
- Consider sector implications: If Tencent’s cost-efficient model scaling proves sustainable, it may intensify competition in the open-source AI space, affecting both AI startups and cloud providers.
- Balance short-term vs. long-term: While near-term earnings growth is muted, the strategic AI investments could support a re-rating of Tencent’s stock if adoption metrics continue to improve.
