New Algorithm Governance Rules Take Effect as Domestic GPUs and AI Large Models Reach Critical Inflection Point

Introduction: Policy and Industry Dual Drivers Send Intensive Signals to the AI Sector

The AI technology sector has recently received a flurry of major developments. From the General Office of the CPC Central Committee and the General Office of the State Council jointly issuing new documents focused on gig worker protections and algorithm governance, to Moore Threads — dubbed "China's first domestic GPU stock" — achieving its first single-quarter net profit turnaround, to DeepSeek signaling model price reductions, the dual threads of policy regulation and industrial development are converging to outline the key trajectory for China's AI industry in the second half of the year.

Algorithm Governance Upgrade: Top Authorities Target Platform Labor Practices and Algorithmic Transparency

The General Office of the CPC Central Committee and the General Office of the State Council have issued a new document focusing on protections for gig economy workers and algorithm governance. The document explicitly requires platform companies to safeguard workers' rights to information and appeal when using algorithms for labor management, task allocation, and performance evaluation, preventing "algorithmic black boxes" from infringing on workers' rights.

This policy did not emerge in a vacuum. As platform economies — including food delivery, ride-hailing, and livestream e-commerce — have boomed, tens of millions of gig workers are now subject to deep algorithmic management. From precise delivery time calculations to automated performance scoring, AI algorithms have improved efficiency while sparking societal debate about workers being "trapped in the system." The release of this high-level document signals that China's governance of AI algorithms is extending from the technical dimension to the institutional level of labor rights protection.

Meanwhile, eight government departments jointly issued new regulations strictly governing online marketing of financial products, further tightening the boundaries of AI applications in the financial sector. Analysts note that from algorithmic labor management to financial marketing, regulators are building a multi-scenario AI governance framework that will have far-reaching implications for the industry.

Moore Threads' Single-Quarter Turnaround: A Breakthrough in Domestic GPU Commercialization

On the industry side, Moore Threads, known as "China's first domestic GPU stock," delivered encouraging news — the company achieved a single-quarter net profit turnaround. This milestone breakthrough signifies that domestic GPU companies are gradually emerging from prolonged losses, with their commercial viability substantively validated.

Domestic GPUs have long faced the dual pressures of technological catch-up and market development. Against the backdrop of international giants like NVIDIA dominating the global AI computing market, domestic manufacturers including Moore Threads have continued to increase R&D investment, striving for breakthroughs in intelligent computing centers, cloud computing, and AI inference scenarios. This turnaround indicates that as demand for domestic AI large model training and inference has surged, real market demand is opening up genuine opportunities for domestic GPUs.

Notably, Hengli Petrochemical, whose subsidiary was placed on the U.S. SDN list, has urgently issued a clarification announcement. The persistent external technology blockades have also objectively accelerated the domestic supply chain's acceptance of homegrown alternatives. Moore Threads' profitability inflection point may well be a landmark event marking the transition of China's domestic AI chip industry from external reliance to self-sustaining growth.

DeepSeek Signals Price Cuts: Ascend Ecosystem Maturing Rapidly

Another development drawing intense industry attention comes from DeepSeek. The company indicated that it expects to significantly reduce prices for its V4-Pro model once Huawei's Ascend 950 super nodes begin shipping in volume in the second half of the year. This statement sends two important signals:

First, Huawei's Ascend AI chip ecosystem is maturing at an accelerating pace. As a next-generation AI training chip, the Ascend 950's super node solution entering mass availability means China's domestic computing infrastructure is about to see a new round of capacity release. DeepSeek, as a leading domestic AI company, proactively linking its product pricing to domestic computing power demonstrates that the Ascend ecosystem's commercial loop is taking shape.

Second, AI large model price wars may intensify further in the second half of the year. As underlying computing costs decline with the scaling of domestic chips, model service providers will inevitably pass cost advantages through to end-user pricing. For the rapidly growing enterprise AI application market, this is undoubtedly a positive signal, but it will also accelerate industry consolidation.

Tesla FSD Entry into China Stalls: Autonomous Driving Deployment Remains Uncertain

In the autonomous driving space, Tesla recently denied rumors that "FSD will go live in China on May 1." Despite Tesla CEO Elon Musk's repeated expressions of intent to bring FSD to the Chinese market, data security, road testing permits, and localization adaptation remain formidable real-world obstacles.

This development also reflects the complexity of deploying cross-border AI technology in the Chinese market. Amid tightening regulations on cross-border data transfers, autonomous driving systems that rely on cloud-based data feedback for model iteration need to find viable localization pathways within the compliance framework. Meanwhile, domestic players such as Huawei, XPeng, and NIO are rapidly advancing their intelligent driving solutions, and the window of opportunity for Tesla's FSD may not remain open for long.

Intensifying International Competition: AI Supply Chain Faces Geopolitical Challenges

On the international front, the EU has included certain Chinese companies on its sanctions list related to Russia, and China's Ministry of Commerce has responded accordingly. While this matter does not directly involve the AI sector, it serves as a warning for Chinese tech companies' global expansion strategies. Against the backdrop of escalating external pressures — including chip export controls and expanding entity lists — self-reliance and controllability across the AI supply chain has become not merely a technological imperative but a strategic one.

Outlook: Three Key Themes for the AI Industry in the Second Half of the Year

Synthesizing the recent intensive stream of policy and industry signals, three key themes merit close attention for China's AI industry development in the second half of the year:

First, the institutionalization of algorithm governance. From top-level design by the CPC Central Committee and State Council to sector-specific regulations by various departments, AI governance is shifting from "principled advocacy" to "rigid constraints," and corporate compliance costs will rise significantly.

Second, the scaled validation of the domestic computing ecosystem. The volume launch of the Ascend 950 and Moore Threads' profitability breakthrough will serve as critical milestones in testing whether domestic AI chips can truly support large-scale commercial deployment.

Third, the deepening and differentiation of large model price wars. As computing costs decline and competition intensifies, AI model service pricing will continue to fall, but vendors with genuine differentiated capabilities are poised to emerge stronger through the industry shakeout.

China's AI industry stands at the crossroads of policy regulation and technological breakthrough. The trajectory of the second half of 2025 may well determine the industry landscape for years to come.