China’s emerging “New New Three” are industrial and service robots, AI large models and innovative drugs—not a formal customs category. The strongest verified example is fruquintinib: developed in China, it received U.S.
Research answer

Create a landscape editorial hero image for this Studio Global article: What is China’s “new new three” export strategy, how does it differ from the “old three” of garments, furniture, and home appliances and the. Article summary: China’s “new new three” is an emerging—not yet a formal customs classification—export narrative: industrial/service robots, AI large models and innovative drugs. It represents a move from exporting manufactured products . Topic tags: general, government, general web, education, academic. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermark
China’s export story is often described in three waves. The “old three” were garments, furniture and household appliances. The “new three” were electric vehicles, lithium-ion batteries and solar products. The emerging “New New Three” are industrial and service robots, AI large models and innovative drugs.36
45
The label is best understood as an economic and policy narrative, not a formal, internationally harmonised customs classification. Its significance is less about a new list of products than about a change in what crosses the border: alongside hardware, China is exporting software, data-enabled services, technical know-how, clinical evidence, licences and complete operating solutions.
| Export wave | Typical products | Main competitive logic | What the buyer receives |
|---|---|---|---|
| Old three | Garments, furniture and home appliances | Labour, scale and cost efficiency | Mostly finished consumer goods |
| New three | Electric vehicles, lithium-ion batteries and solar products | Capital-intensive manufacturing and supply-chain scale | Physical green hardware and production capacity |
| New New Three | Robots, AI models and innovative drugs | Research, software, data, IP and continuous iteration | Hardware plus services, digital capability or a licence and regulatory package |
The first two waves are primarily about moving goods. The latest trio is more often sold as a continuing relationship: a robot integrated into a workplace, an AI model delivered through an API or cloud deployment, or a drug candidate licensed for overseas development and commercialisation. That is why proponents describe the shift as moving from “Made in China” toward “Created in China.”36
44
45
An industrial or service robot is rarely valuable as a standalone machine. Its commercial package can include sensors, control software, integration with existing equipment, maintenance, spare parts, remote diagnostics, safety documentation and adaptation to a customer’s workflow.
That matters especially when robots perform dangerous or difficult work. Reports have highlighted Chinese wall-climbing cleaning robots in Australia as an example of technology aimed at high-rise maintenance rather than ordinary household automation.25 The report demonstrates the intended export model, but the available evidence does not establish the deployment’s scale, vendor, commercial performance or safety record.
Trade reporting provides a stronger indication of sector momentum: China’s industrial-robot exports in the first half of 2026 were reported at ¥6.29 billion, up 18.6% year over year, with products reaching 141 countries and regions. Surgical-robot exports were reported at ¥480 million.39
47 These figures should not be read as proof that every type of robotics is globally competitive; the “New New Three” remains a broad narrative covering very different markets.
AI exports can cross borders without a conventional shipment. They may take the form of an open-weight model, an API, a cloud service, a customised enterprise system, an industry-specific algorithm or an implementation and support contract.
The differentiator is therefore not simply a benchmark score. A model must work in a local language, fit the customer’s data environment, connect to existing software and meet requirements for security, auditability and governance. In industrial settings, the export may be the model together with the infrastructure and operational process needed to use it.
A reported example from Brazil describes the Bright Power Large Model, developed by State Grid, being used in a transmission operations-and-maintenance platform. The stated use case focuses on helping inspect transmission lines across difficult terrain where helicopter-based manual inspections can be inefficient.26 This is evidence of a proposed or reported operational application, not enough evidence to measure its reliability, autonomy or commercial scale.
Brazil’s broader AI strategy also illustrates the geopolitical complexity. Reuters reported in August 2026 that Brazil planned to divide AI investments between Chinese and U.S. partners, including a supercomputing project involving Huawei and iFlytek.22 For Chinese suppliers, that creates market access; for the customer, it also reflects an effort to avoid relying exclusively on either power.
Pharmaceutical exports often do not look like containers of finished medicines. They can involve patents, out-licensing, co-development, clinical-trial data, manufacturing know-how, regulatory submissions and overseas commercialisation.
Fruquintinib—marketed in China as Elunate and outside mainland China, Hong Kong and Macau as FRUZAQLA—shows how this model works. HUTCHMED developed and marketed the drug in China, while Takeda holds the exclusive licence to further develop, manufacture and commercialise it outside those Chinese markets.13
15
The drug was approved in China in 2018, by the U.S. FDA on November 8, 2023, by the European Commission in June 2024 and in Japan in September 2024.1
3
13 The FDA says its approval relied primarily on two clinical trials enrolling 1,107 patients; one was conducted entirely in China and the other was multinational.
3
That makes fruquintinib a meaningful example of China-originated pharmaceutical innovation clearing major overseas regulatory hurdles. It does not, by itself, show that Chinese innovative drugs have achieved broad global commercial acceptance. The example also demonstrates the continuing role of an established multinational partner in international development and market access.
The economic opportunity is that more of the value may remain in intellectual property and recurring services rather than in a one-time sale.
This also changes the competitive objective. A company is not only trying to sell a product; it is trying to become part of a customer’s workflow, technology stack or drug-development pathway. The available sources support the broader shift toward knowledge- and technology-intensive exports, but they do not establish that Chinese AI interfaces or robotics standards have become global defaults.44
45
The new model can reduce reliance on Western finished products without eliminating international dependencies. High-end robotics may require specialised components, software and certification expertise. AI systems remain connected to compute, chips, cloud infrastructure, data governance and cybersecurity. Drug developers may need overseas partners for clinical development, regulatory execution, manufacturing validation, reimbursement and distribution.
Fruquintinib makes this mixed picture clear: Chinese-origin drug discovery is central to the product’s story, while Takeda’s exclusive ex-China licence is central to its global development and commercialisation.13
15 The lesson is not that China is either fully independent or merely dependent. It is that high-value exports are assembled through cross-border ecosystems, even when the underlying innovation originates in China.
The three sectors face different versions of the same problem: technical capability must be converted into trust.
Buyers in sensitive or hazardous environments will need confidence in worker safety, product liability, remote access, cybersecurity, data handling, spare-parts continuity and long-term maintenance. A robot that works in a demonstration is not automatically ready for a regulated workplace or critical facility.
AI supplied to utilities, public agencies or other critical infrastructure raises questions beyond accuracy. Customers may need assurances about data localisation, cross-border transfers, audit trails, model updates, procurement rules, privacy, cybersecurity and responsibility for errors. The Brazilian case shows the strategic appeal of Chinese AI infrastructure, while the split between Chinese and U.S. partners shows that geopolitical risk remains part of the purchasing decision.22
26
Medicines must still clear jurisdiction-specific requirements for quality, safety and efficacy, as well as inspections, pharmacovigilance, labelling and reimbursement processes. The FDA and European approvals of FRUZAQLA show that the pathway is possible; they do not remove the cost, time or uncertainty of global pharmaceutical development.2
3
4
The “New New Three” will be judged less by the novelty of its label than by whether overseas customers accept five conditions:
China’s export progression is therefore not a simple replacement of one trio by another. Garments, furniture and appliances established manufacturing scale. Electric vehicles, batteries and solar products extended that scale into capital-intensive green industries. Robots, AI and innovative drugs aim to export something harder to copy: Chinese-origin intellectual property embedded in systems, services and regulated innovation.
The opportunity is higher-value, recurring international revenue. The constraint is equally clear: every sale becomes a test of standards compatibility, regulatory credibility, supply-chain resilience and global trust.
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
China’s emerging “New New Three” are industrial and service robots, AI large models and innovative drugs—not a formal customs category.
China’s emerging “New New Three” are industrial and service robots, AI large models and innovative drugs—not a formal customs category. The strongest verified example is fruquintinib: developed in China, it received U.S.
The model is promising but not settled: robot and AI deployments still face questions about safety, cybersecurity, data, interoperability, supply chains and political trust.