How can algorithm technology collaboration with Chinese enterprises proceed under export restriction policies?

Currently, the competitive landscape of the global artificial intelligence sector is being reshaped by increasingly stringent export control policies. These policies aim to regulate the flow of cutting-edge algorithms and computing power technologies, casting a shadow over global technological collaboration. However, China's vast market, diverse application scenarios, and strong engineering capabilities make it an indispensable part of the global AI ecosystem. Against this backdrop, a critical question confronts international companies: how can they identify viable pathways for collaborating with Chinese enterprises on algorithm technology while adhering to compliance bottom lines?

The answer is not "no cooperation," but rather a shift from the previously extensive model of technology transfer to more refined, flexible, and innovative collaboration models. The key to breaking the impasse lies in deeply understanding the boundaries of restrictions and maximizing the value of cooperation within those boundaries.

I. Strategic Shift: From "Technology Transfer" to "Value Co-creation"

Traditional models involving the exchange or transfer of underlying core technologies now carry extremely high risks. Successful cooperation must undergo a strategic shift, focusing on areas that do not directly touch controlled areas yet can create significant commercial value. The safest path is to avoid transfers involving algorithm ownership—particularly underlying model weights and source code. Instead, models like Technology Licensing or Algorithm-as-a-Service (AaaS) should be adopted. International companies can package their algorithms as APIs or SDKs and license their use to Chinese partners, rather than selling them outright. This approach is similar to the model supporting TikTok, where core intellectual property, after compliance review, is output as a service, with returns realized through licensing fees. This indicates that substantial cooperation space remains in application-specific algorithms for particular vertical industries. For example:

Industrial AI: Collaboratively develop algorithms for visual inspection, predictive maintenance, and process optimization in smart manufacturing. These algorithms are closely tailored to specific equipment and processes, have lower generality, and are thus less sensitive.

Medical AI: Jointly research algorithms for medical image analysis, drug development assistance, etc., leveraging China's advantages in clinical data for model optimization, while ensuring the core training framework and data processing comply with ethics and regulations.

Consumer Internet: Conduct joint experimentation and optimization in areas like recommendation systems, advertising algorithms, and user experience enhancement to collectively improve services for the vast Chinese market.

II. Compliance First: Building a "Navigation System" to Traverse Risk Zones

Before initiating cooperation, it is essential to conduct strict screening of every technology and data flow involved in the collaboration with specialized legal teams to determine if they fall under export control lists. This requires establishing a dynamic "compliance calendar" to continuously track policy changes in China, the US, and other relevant jurisdictions. Feasible solutions include:

Data Localization: Establish fully independent data centers within China, where all data processing and analysis serving the Chinese market are completed domestically. Only anonymized analysis results or model increments—not raw data—are used for global model improvements after obtaining necessary permissions.

Third-party Supervision Model: Learn from successful cases by introducing internationally recognized third-party institutions to supervise data storage and processing, ensuring data security and compliant usage.

III. Ecosystem Integration: Leveraging China's Domestic Technological Wave

In response to restrictions, China is fully committed to building a self-sufficient and controllable technology ecosystem. International companies can embrace this trend and turn it into a cooperation opportunity.

Adapting to Domestic Computing Platforms: Actively adapt to domestic AI chips and their supporting software stacks, such as Huawei's Ascend and Hygon. This not only mitigates supply chain risks but also demonstrates a commitment to long-term cooperation to the Chinese market, thereby unlocking more opportunities.

Utilizing China's Mature AI Platforms: Many Chinese tech companies offer robust open-source frameworks or commercial AI platforms. Developing upper-layer applications based on these platforms can effectively reduce technical barriers and compliance risks.

Exploring Third-party Markets: Expand the scope of cooperation beyond China. Combine the global perspective and brand influence of international companies with the cost control and engineering implementation capabilities of Chinese enterprises to jointly explore business opportunities in emerging markets along the "Belt and Road," Southeast Asia, the Middle East, and elsewhere, achieving win-win outcomes.

IV. Potential Risks and Countermeasures

The greatest risk stems from the inherent uncertainty of the policies themselves. Companies need to establish agile organizational structures and prepare multiple contingency plans (Plan Bs) to switch quickly in case of sudden policy changes.

Intellectual Property Protection: Core IP must be protected through rigorous contractual terms and technical measures (e.g., code obfuscation, modular licensing). Consider keeping the most critical parts of the algorithm offshore, only exporting interfaces and results.

Geopolitical Pressure: Cooperation may attract regulatory scrutiny or public opinion pressure in other markets. Companies need to communicate effectively, transparently demonstrating the compliance and commercial rationale of the cooperation.

Export control policies have not closed the door to cooperation but have instead defined a narrower, more skill-intensive track. The core of collaboration will shift from "what technology can you give me" to "what value can we create together." By pivoting towards applied innovation, building a rock-solid compliance system, and actively integrating into the local ecosystem, international companies can still find their place and achieve resilient growth in China, one of the world's most crucial AI battlegrounds.