Deloitte China Monthly Report Issue 102

Economy

A Surging RMB Paves the Way for a Rate Cut

The Chinese economy in 2025 could be characterized as a world of ice and fire — two extremes depicting resilient exports and subdued domestic demand. Data from December 2025 reinforced this dichotomy: exports grew 6.6% year-on-year, sustaining November’s impressive 5.9% increase and bringing full-year growth to 5.5%. Such a stellar performance is no small feat, given that exports have consistently driven growth since the COVID-19 outbreak. Moreover, China’s exports have faced numerous headwinds — from tariffs to protectionist measures — for almost eight years.

It would be simplistic to attribute weak domestic demand solely to the property sector’s woes. Nonetheless, the sector’s excesses could take years to unwind, particularly given the severe supply-demand imbalance in urban China. The floor space of commercial buildings for sale, for instance, has risen from 500 million square meters in 2020 to 770 million square meters in 2025. Property investment in December fell 17.2% year-on-year, extending November’s 15.9% decline. Existing home prices in most cities dropped around 5% year-on-year in January 2026. The property sector’s sharp correction has weighed heavily on retail sales — a proxy for urban consumption. The question now is whether this trend will persist in 2026.

The Property Sector: Outlook Still Clouded

The outlook for the property sector remains uncertain. S&P has forecast another year of double-digit declines in sales. More notably, the rating agency revised its original projection of a 5%–8% drop in 2026 sales to a steeper decline of 10%–14%. This adjustment likely reflects both a larger-than-expected inventory overhang and the perception that policymakers are reluctant to bail out major developers.

However, the latest issue of Caijing offers a somewhat brighter picture. An investigative report on the real estate sector (source: https://m.caijing.com.cn/article/202602/399182) highlighted a surge in transaction volumes in many cities — often a leading indicator for price recovery. One thing is clear: the property sector, which accounts for roughly a quarter of economic activity, will continue to exert outsized influence on investment, consumption, and trade (through reduced import demand).

What more can be done to stabilize the housing market? Qiushi’s editorial of January 1, 2026 emphasized the “financial nature of housing,” signaling a policy shift toward encouraging homeownership — a positive sign. Yet, recent stock market trends suggest investors remain unconvinced about stronger policy support for highly leveraged developers such as Vanke. While local governments have tried to reduce inventories, the central government could better support both struggling developers and homebuyers facing mortgage difficulties.

Chart: Languishing stock prices of major developers have highlighted investors’ desires for  stronger policy support 

Source: Wind

If Caijing’s report truly signals light at the end of the tunnel, policymakers may remain cautious. But if S&P’s revised forecast proves more accurate, additional policy support will likely follow — although a policy “bazooka” seems improbable. The risk is that such restraint could exacerbate deflationary pressures.

Growth Target and Policy Considerations

With the annual “Two Sessions” approaching, the key question is the 2026 growth target. We expect it to be set at “around 5%,” as in 2025, but with a stronger emphasis on  high-quality growth, allowing more flexibility in actual GDP figures. Our base case is 4.5%, primarily because the housing market will remain a drag — though perhaps less severe than S&P’s forecast.

Can Export Strength Continue?

A bigger question is whether China’s robust export momentum can continue through 2026. From a competitiveness standpoint, China remains formidable, even amid an appreciating RMB. However, if China’s export surge causes economic or social disruptions abroad, its major trading partners may adopt defensive measures. In ASEAN, where China plays an increasingly important role as both a leading FDI provider and a stabilizing force, some backlash is inevitable, though broad tariff hikes seem unlikely. Over time, Chinese companies operating in ASEAN must localize more rapidly — especially if their competitive products are perceived as threatening industrialization (as in Indonesia).

In developed markets, the situation is more complex. Currently, average U.S. tariffs on Chinese goods hover around 30% — less punitive than in early 2025 but still high enough to encourage Chinese exporters to reroute shipments to the U.S. and other developed markets. (UK steel industry has ‘two months to be saved’, warns Tata Steel | The Standard)

Chart: China’s shifting value chain across overseas markets

Source: Wind

Will President Trump’s scheduled visit to China in April bring about a more lasting trade deal? And if so, will tariffs be lowered? Thanks to China’s diversification efforts, bilateral trade imbalances have become less of a U.S. concern. However, with inflation and affordability dominating domestic debates ahead of the midterms, these issues could still surface prominently. Rumors of China reducing U.S. Treasury holdings may also arise as a negotiating topic.

Could China help lead an appreciation of other surplus economies’ currencies against the dollar in exchange for eased investment restrictions in the U.S.? Unlike 2018, today’s bilateral relationship centers on competition in high-tech sectors while avoiding overt escalation. Thus, reducing the trade deficit through greater purchases of U.S. agricultural or energy products may serve only as low-hanging fruit — as could a minor tariff cut on fentanyl-related items. A more meaningful deal would require institutionalized communication and policy coordination, even if contentious geopolitical issues are sidestepped.

Such an agreement would likely cause trade diversion. Accordingly, pressure for China to narrow its trade imbalances may also come from other partners — notably Europe. China’s trade surplus has ballooned with both the EU and the UK. Germany, historically running a surplus with China, now faces competitive pressure from Chinese exporters both at home and abroad. Europe’s likely response is more assertive industrial policy.

What does this mean for China? Brussels will seek wider market access while pressing Beijing to curb industrial overcapacity. In late February, German Chancellor Friedrich Merz made his first visit to China since taking office.

For China, diplomatic stability could draw renewed European investment — as seen in AstraZeneca's recent plan to invest USD 15 billion in China. Meanwhile, manufacturing in Germany and France remains far more strategic than in the UK. French President has urged Europe to take "urgent action and concrete decisions" by June amid intensifying competition.

For China, its ongoing anti-involution campaign will likely involve more judicious industrial policy and deemphasizing headline GDP growth. If China and the EU can find common ground on market access, the long-stalled Comprehensive Agreement on Investment (CAI) may see new momentum. Still, a deeper Sino-U.S. trade accord in 2026 would almost certainly trigger some degree of trade diversion — with China increasing purchases of U.S. goods at the expense of others. Overall, China’s exports may remain strong, but backlash against its widening trade surplus will likely become more pronounced.

Policy Outlook: Monetary Easing Ahead?

If a fiscal “bazooka” isn’t in the pipeline, a significant interest rate cut could serve as a viable policy alternative. Qiushi’s recent article calling to “proactively push for a reasonable rebound in the general price level” is a clear endorsement of reflation and a warning against deflation. A strengthening RMB, paradoxically, makes a near-term rate cut more likely. China’s 2026 growth target appears attainable — but genuine reflation is by no means assured.

TMT

"Physical AI" Reshaping the Real World

At the 2026 Consumer Electronics Show, NVIDIA hailed Physical AI as the next "ChatGPT moment" for artificial intelligence. Unlike generative AI, Physical AI directly collects and processes data from the real world to make decisions and take action. By integrating large AI models, physics engines, and multidisciplinary simulation, it achieves high-precision dynamic simulation, prediction, and optimization of the real world, empowering traditional industries. Its applications are particularly prominent in robotics and automation. If AI represented by ChatGPT has reshaped knowledge work and content creation, then Physical AI-centered systems will transform every physical aspect of manufacturing, logistics, transportation, and even services. This is not only crucial for the future competitiveness, supply chain resilience, and strategic autonomy of national manufacturing but also represents a new high ground in global technology and industrial competition.

I. Physical AI Leverages a Trillion-Dollar Market
Physical AI’s trillion-dollar market potential stems directly from the efficiency gains and value redistribution within existing trillion-dollar industries like manufacturing, transportation, logistics, and healthcare. In the short term, Physical AI will drive robots to evolve from mere executors into intelligent agents with environmental perception and autonomous decision-making capabilities. In the medium to long term, domestic manufacturers possess advantages for rapid iteration within local application scenarios. If breakthroughs in domestic chips and large models continue, Chinese robot manufacturers and related supply chain companies are poised for exponential growth in the Physical AI implementation phase.

II. Leading Overseas Players Actively Deploy, Domestic Industrial Chain Takes Shape
Overseas, Physical AI is already applied in areas like embodied intelligence. NVIDIA, leveraging its Omniverse simulation capabilities, has built a comprehensive ecosystem covering hardware, algorithms, development tools, and applications. Its "Cosmos" model described as the world's first “World Model”. Google is also actively advancing its Physical AI strategy, extending artificial general intelligence from digital to the real physical world. Domestically, the Physical AI industrial chain is rapidly takin ghsape, building a full-stack ecosystem from underlying perception to upper-layer applications. The pathway "Simulation - Perception - Robot" reveals the key links. For example, SoCrane Tech's "Tiangong Kaiwu Platform"—based on generative Physical AI and real-scene rendering technology—enables multi-physics design, simulation, and optimization under four-dimensional spatiotemporal coupling. It can quickly build real scenes and simulate physical field interactions, while also validating applications in scenarios such as robotics and intelligent driving through virtual training, providing an integrated Physical AI toolkit. This capability to deeply integrate computational physics with AI signifies that domestic enterprises already possess the potential to compete with international giants in the foundational platform domain of Physical AI.

III. Physical AI Competitiveness is Rooted in Breakthroughs in Advanced Chips and Computing Power
Currently, computing infrastructure represented by NVIDIA's AI factories and Ascend super-node clusters provides high-performance intelligent computing power to support Physical AI model training. Chip capabilities continue to strengthen globally. NVIDIA's GPU series has a clear development roadmap, while ASIC chips from leaders like Google, Amazon, as well as domestic players like Huawei and Cambricon are iterating constantly, forming the hardware foundation for Physical AI development. In the future, more advanced accelerated computing chips will be key to overcoming computing bottlenecks, enabling large-scale, high-precision, real-time simulation and training of complex physical scenarios.

IV. The Application Scope of Physical AI is Expected to Expand Continuously
This includes design optimization in high-end manufacturing, simulation testing for automotive intelligent driving, drug design and development in pharmaceuticals, precise climate prediction in meteorology, and novel particle discovery in fundamental science. Its application is gradually expanding from traditional manufacturing to more frontier scientific fields. Physical AI solutions will continue to improve, delivering high precision, high efficiency, and low-cost benefits for clients. NVIDIA for example, has open-sourced Physical AI models for engineering physics, biomedicine and robotics, thereby supporting scalable, high-performance scientific AI model development. These models are now widely used on open-source platforms and are expected to gradually form a comprehensive platform covering physical models, AI frameworks, datasets, and pre-trained model libraries, significantly lowering the barriers to development for Physical AI. In turn, this will promote its application in niche areas, continuously enhancing the reliability and replicability of solutions.

V. Unleashing Physical AI's Potential Still Requires Overcoming Major Challenges
Despite its transformative promise, Physical AI still faces several fundamental challenges. First, building a comprehensive “world mode” requires huge volumes of data, making it imperative to find more economical and energy-efficient data processing methods than those used in traditional machine learning. Second, as Physical AI performs tasks directly in the physical environment and must respond in real time, powerful, distributed computing capabilities embedded on the devices themselves are essential. Therefore, further development of micro-computing and connectivity technologies is crucial. Third, the development of robotic devices driven by Physical AI typically involves long cycles and high costs. Fourth, traditional robots and automation equipment ensure safety through rigorous inspection, verification, and validation during design, development, and debugging. However, these methods cannot be directly applied to AI systems, which are characterized by limited transparency, susceptibility to errors, bias, and unpredictability.

The transformation led by Physical AI will likely surpass the mobile internet era in both depth and breadth. China possesses the world's richest manufacturing scenarios and market demand. To gain the initiative in this competition, advancement can be accelerated across the following levels:

Strengthen the foundational layer: Ramp up indigenous R&D in core hardware and software such as high-end sensors, dedicated AI chips, and real-time operating systems to reinforce the technological base.

Build the platform layer: Promote the establishment of open, collaborative Physical AI development platforms and ecosystems to lower the barriers to adoption and foster resource sharing and collaborative innovation.

Deepen the application layer: Create a batch of replicable, scalable benchmark applications in advantageous fields like new energy vehicles, high-end equipment, and smart logistics, driving continuous technological iteration and industrial maturity through real-world scenarios.

Automotive

Finding a Way Out Amidst a Siege of Tariffs

Chinese automakers entered 2026 facing a global trade environment marked by significant divergence and rapid shifts, presenting new challenges and opportunities for their globalization strategies. On one hand, the European Union—previously a stronghold of high tariffs—has made a strategic adjustment. In January 2026, it reached an important framework agreement with China, agreeing to replace anti-subsidy tariffs that previously reached as high as 45.3% with a “minimum price commitment” mechanism. Canada, at the start of the year, announced a substantial reduction in its additional tariff rate on imported Chinese EVs from 100% to 6.1%, while imposing quota limits. On the other hand, emerging economies such as Mexico, Brazil, Indonesia, and Malaysia have collectively shifted toward “conditional market access,” significantly raising tariffs or introducing strict localization requirements. These changes indicate that Chinese automakers must adopt flexible, country-specific market strategies. Their globalization model must also evolve from a previous reliance on whole-vehicle exports and KD (Knock-Down) assembly to a new phase of deep localization and full value-chain operations.

Shifting Trade Winds: Mature Markets Lower Entry Barriers

On January 16, 2026, Canada announced the cancellation of its previous 100% tariff on Chinese electric vehicles (including pure electric and hybrid passenger and commercial vehicles), replacing it with a preferential tariff rate of 6.1%. It also set an initial import quota of 49,000 vehicles per year (scheduled to increase to 70,000 within the next five years), with over half of imported vehicles required to be affordable models priced at CAD 35,000 (approximately USD 25,000) or below. Although Canada’s annual new energy vehicle (NEV) sales volume remains relatively modest at fewer than 180,000 units, the nearly 50,000-vehicle quota accounts for about 30% of local NEV sales, making it a strategically important market for Chinese automakers seeking a foothold in mature economies.

On January 12, 2026, the European Commission issued the “Guidance Document on Submitting Price Undertaking Applications,” allowing Chinese electric vehicle manufacturers (excluding PHEVs) to replace the previous 7.8%–35.3% anti-subsidy tariffs by submitting minimum price commitments. Manufacturers must set a minimum selling price for each battery electric vehicle (BEV) exported to Europe, ensuring that the effect is equivalent to the punitive tariffs.

Contraction of Policy Benefits in Emerging Markets, Stricter Access

Mexico and Brazil, the first and second largest export destinations for Chinese automakers, have both adjusted their tariff policies on imported Chinese vehicles. Starting in 2026, Mexico imposed a higher tariff of 50% (up from 20%) on EVs from countries without a free trade agreement, including China, South Korea, and India. Brazil reinstated import tariffs on new energy vehicles beginning in January 2024, with rates expected to uniformly rise to 35% by July 2026. Furthermore, Brazil plans to impose a 35% tariff on EVs and hybrid vehicles imported in CKD/SKD form starting January 2027—18 months earlier than originally scheduled.

Beginning in 2026, Indonesia will completely suspend duty-free treatment for whole-vehicle imports and provide a two-year grace period for companies that have already utilized import quotas. During this period, automakers must produce an equivalent number of locally manufactured vehicles based on historical import volumes, meeting localization requirements of 40% in 2026 and 60% in 2027. Malaysia has announced that starting in 2026, it will reinstate import duties, excise taxes, and road taxes on fully imported EVs, while vehicles assembled locally (CKD) will continue to enjoy tax incentives until the end of 2027.

Faced with these dynamic changes in the global tariff environment, Chinese automakers can respond through several strategic adjustments:

Transform tariff and policy risks into differentiated competitive advantages. Under the EU’s “minimum price mechanism,” automakers must make careful strategic trade-offs. On one hand, by committing to a minimum price, they can retain what would otherwise be paid as countervailing duties to the EU as internal corporate profit, providing funding for market expansion and R&D—particularly for manufacturers subject to higher tariff rates. In this sense, joining the MIP framework allows firms to convert tax liabilities into retained earnings. On the other hand, imposing a price floor weakens the competitiveness of low-end models, forcing automakers to shift exports toward intelligent, high-specification, premium-priced vehicles. The strategy of “adding features without raising prices” also faces constraints, as EU authorities may assess whether vehicle specifications align with declared prices, potentially triggering a re-evaluation of the minimum price commitment. Automakers must therefore increasingly rely on technology- and service-driven premium value, such as charging solutions, insurance, extended warranties, and bundled services, to support upmarket positioning. At the same time, the MIP mechanism constrains pricing flexibility and exposes firms to heightened compliance and audit risks.

Upgrade from SKD to CKD and local production models to comprehensively raise localization levels. Compared with fully integrated manufacturing plants—which require higher upfront investment and longer payback periods—KD factories have become the mainstream choice for Chinese automakers establishing overseas operations. However, simple assembly plants are no longer sufficient to meet evolving industrial policy requirements. For example, Brazil has accelerated tariff increases on pure electric and plug-in hybrid vehicles exported in knockdown form, while Indonesia requires local factories to achieve at least a 40% localization rate to qualify for government procurement priority and VAT incentives. These policy shifts indicate that the model of importing major assemblies from China and conducting only final assembly locally will face significant cost restructuring. Automakers must accelerate localization by embedding themselves more deeply into regional industrial chains. Key measures include upgrading from Semi Knock-Down (SKD) to Complete Knock-Down (CKD) assembly; investing in welding, painting, and final assembly workshops; prioritizing cooperation with local suppliers or Chinese suppliers with overseas production bases—particularly in core components such as batteries, motors, and chassis; and hiring local engineers and technical workers to generate employment and industrial added value.

Enhance the “tariff elasticity” of regional production capacity. Chinese automakers can also leverage free trade agreement networks and tariff differentials across regions to dynamically allocate global production capacity, thereby mitigating tariff exposure. The export of vehicles from Thailand-based factories to the EU offers a notable example of this approach. Companies such as BYD and Changan export pure electric models produced in their Thai plants to the EU, avoiding anti-subsidy tariffs applied to China-origin vehicles while benefiting from preferential tariff arrangements between Thailand and the EU.