AI Regulation Blunder Highlights Value of Sovereign AI; Bullish Targets Set for Chinese ADRs

Overview

Using Anthropic's suspension of its latest AI model, Claude Fable 5, as a starting point, this report delves into the geopolitical risks in AI development and their impact on the industry landscape. The report argues that while this event may be temporary, it exposes the risks of over-reliance on single suppliers—especially cloud providers subject to specific national policies—thereby providing a strong case for 'Sovereign AI' and localized deployment. Meanwhile, the report notes that Chinese AI labs are making rapid technological progress, gradually narrowing the gap with top global standards. On the investment front, the report maintains Outperform ratings on Tencent Holdings and Alibaba Group, setting high target prices for both, suggesting that despite macro and regulatory risks, valuations for these two companies remain attractive.

Core views

The core views of the report focus on AI geopolitical risks and valuation opportunities in Chinese tech stocks. First, regarding AI development and geopolitics, the report points out that Anthropic's suspension of the Claude Fable 5 model represents the first 'regression' in the AI frontier. This intervention by the U.S. government on national security grounds has triggered widespread discussion within the developer community. The report argues that as Agentic AI capabilities improve, these tools will become embedded in critical workflows. If access is controlled by third parties and could be cut off at any time, large enterprises and non-U.S. sovereign operators will question the reliability of cloud-deployed AI models. This constitutes a strong driver for the development of 'Sovereign AI' (i.e., countries or large enterprises building or controlling their own AI infrastructure). Second, against the backdrop of geopolitical competition, the performance of Chinese AI labs warrants attention. Over the weekend, leading Chinese AI labs Kimi and Z.ai released new models, K2.7 Code and GLM-5.2, respectively. Both online developer feedback and local benchmark tests by the report's authors indicate positive progress in reasoning efficiency and capabilities. This suggests that leading Chinese labs have the ability to keep pace globally and may even seize the moral high ground in the open-weights model space. Finally, regarding specific targets, the report believes the U.S. government's decision to restrict access to frontier models based on nationality reminds investors of the intersection between (geo)political narratives and AI development. For top U.S. AI labs, this is not favorable for use cases relying on reliability and may also affect market narratives for subsequent IPOs. In contrast, Chinese tech giants face a different regulatory environment. The report maintains a bullish view on Tencent and Alibaba, believing their valuations fully reflect risks and offer allocation value. In terms of valuation, Tencent's target price is set at HKD 780 based on 20x FY+1 P/E; Alibaba's target price is set at USD 180 (or HKD 176) based on Sum-of-the-Parts (SOTP) valuation, incorporating FY+1 revenue and profits from core e-commerce and cloud businesses.

Analysis framework

The institution's analytical approach adopts a progressive method moving from macro industry trends to micro company valuations. The first step is event-driven industry logic deduction. By interpreting the specific event of Anthropic's model suspension, analysts distilled two core concepts: 'AI supply chain security' and 'Sovereign AI.' This analytical method emphasizes potential structural changes in business models and user behavior within the tech industry when sudden shifts occur in non-technical fields (such as national security policy). The second step is horizontal comparison and competitive landscape analysis. After discussing U.S. regulatory actions, analysts immediately introduced developments from Chinese competitors (new model releases by Kimi and Z.ai). By comparing progress in the AI frontier between China and the U.S., they demonstrated the resilience and potential of Chinese technology during its catch-up phase, thereby providing fundamental support for a bullish stance on related Chinese assets. The third step involves traditional relative valuation and sum-of-the-parts valuation. Having established an industry preference, analysts returned to specific financial models. For Tencent, a relatively simple P/E multiple method was used, selecting FY+1 (next four quarters) as the baseline and applying a 20x P/E multiple, reflecting recognition of stable cash flows from gaming and advertising businesses. For Alibaba, a more complex SOTP (Sum of the Parts) method was adopted to separately assess profit contributions from core e-commerce and cloud computing businesses; this method better reflects the true value of a diversified conglomerate, especially when growth prospects vary significantly across business segments.

Methodology notes

• Event Arbitrage & Behavioral FinanceEvent-driven analysis

  Event-Driven Analysis

  By analyzing the specific event of Anthropic's model suspension, the report deduces its long-term impact on trust structures and procurement strategies across the entire AI industry, rather than focusing solely on short-term stock price fluctuations.

• Valuation MethodologySOTP Valuation

  SOTP Valuation

  Used for Alibaba valuation. The company is split into different business segments such as core e-commerce and cloud; reasonable valuations are assigned to each and then summed up, suitable for large groups with diverse businesses and varying maturity levels.

• Valuation MethodologyPE/PEG valuation

  PE Valuation

  Used for Tencent valuation. The target price is determined by applying a reasonable P/E multiple to expected earnings per share (EPS) for the next year, a common method for evaluating mature, profitable internet companies.

• Industry/Sector Analysis FrameworkSubstitution Effect Analysis

  Sovereign AI Substitution Effect

  The report implies that due to concerns over the reliability of U.S. cloud AI models, enterprises and governments in non-U.S. markets may shift towards self-built or localized AI solutions, creating substitution pressure on existing global cloud AI giants.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Tencent Holdings (700.HK)

  Beneficiary/Core Coverage

  Strengths

  Strong gaming and advertising businesses, stable cash flow, reasonable valuation (20x FY+1 PE).

  Weaknesses

  Exposed to macroeconomic volatility, user engagement fluctuations, and antitrust regulatory risks.

  Comparison

  Compared to other Chinese internet platforms, Tencent has deeper moats in social media and gaming.

  Risks

  Macro credit risk, weak retail consumption, competitive threats.

• Alibaba Group (BABA.US / 9988.HK)

  Beneficiary/Core Coverage

  Strengths

  Solid foundation in core e-commerce and cloud businesses; SOTP valuation indicates undervaluation; innovative businesses show losses but possess significant long-term potential.

  Weaknesses

  Fluctuating user engagement on Taobao/Tmall, intense competition, innovative businesses dragging down overall profits.

  Comparison

  Compared to pure-play e-commerce platforms, Alibaba possesses diversified infrastructure advantages including cloud services and logistics.

  Risks

  Macro risks, regulatory risks, intensified market competition, continued losses in innovative businesses.

Key data

• Tencent Target Price780 HKDImplies ~49.9% potential upside, based on 20x FY+1 PE
• Alibaba (US) Target Price180 USDImplies ~26.2% potential upside, based on SOTP valuation
• Alibaba (HK) Target Price176 HKDImplies ~42.4% potential upside, based on SOTP valuation
• Tencent FY+1 Est. EPS30.00 CNYBaseline EPS used to calculate target price
• Alibaba FY+1 Est. EPS26.82 CNY (US) / 3.69 HKD (HK)Baseline EPS used to calculate target price

Impact & implications

The report views the Anthropic incident as a significant signal indicating that the global deployment of AI technology is being profoundly reshaped by geopolitical factors. For investors, this means re-evaluating risk exposure for enterprises heavily reliant on single U.S. cloud AI suppliers. Simultaneously, this creates potential structural opportunities for domestic Chinese AI infrastructure providers and application-layer companies, as rising demand for 'Sovereign AI' will foster local ecosystem development. For Tencent and Alibaba, despite macroeconomic and regulatory uncertainties, their robust cash flow generation capabilities and dominant positions in their respective sectors provide a high margin of safety and investment attractiveness at current valuation levels.

Risks

• Macroeconomic risks (e.g., credit tightening, weak retail consumption)
• Fluctuations in platform user engagement
• Intense competition from other internet platforms
• Regulatory risks (particularly antitrust regulations in China)
• Continued losses in Alibaba's innovation and other business segments

What to watch

• Status of Anthropic model restoration and subsequent U.S. government policies on AI export controls
• Practical application results and market feedback for new models from Chinese AI labs
• User activity and advertising revenue growth for Tencent and Alibaba
• Progress in profitability improvement for Alibaba Cloud