Chinese autos and Asia emerging robotics: China autos are navigating a weak domestic cycle through exports, technology and consolidation, while robotics commercialization remains selective and gradual.
Bernstein's China EV and robotics tour finds domestic auto demand and profitability under pressure, but sees exports, vertical integration, electrification and industry consolidation as important offsets. The report is constructive on selected OEM and robotics names while stressing policy, competition and execution risks.
Summary
Bernstein's China EV and robotics tour finds domestic auto demand and profitability under pressure, but sees exports, vertical integration, electrification and industry consolidation as important offsets. The report is constructive on selected OEM and robotics names while stressing policy, competition and execution risks.
- China's 2026 eight-month wholesale auto volumes fell 5.3% year on year as domestic demand declined 19.7%, while exports rose 71.8%.
- Participants expect EV penetration to reach 80-90% by 2030, but domestic demand in 2027 is viewed cautiously.
- OEMs are increasing in-house battery-pack and ADAS-chip development to reduce costs and improve supply security.
- Humanoid-robot mass adoption is generally viewed as three to five years away, although XPeng targets early production in 2027.
- Government-led capacity rationalization and improved supplier-payment practices could support industry sustainability.
- Bernstein rates BYD, Geely and Xiaomi Outperform among Chinese autos; Hesai, Tuopu and Shuanghuan are Outperform in Asia emerging robotics.
Report Interpretation
Overview
The report summarizes Bernstein's China Humanoid Robotics & Auto Supply Chain Tour and China EV & Battery Value Chain Tour. Its central conclusion is that China autos remain in a difficult domestic demand and profitability environment, but leading companies are building resilience through exports, technology ownership, vertical integration and consolidation; emerging robotics offers longer-term growth but uneven and execution-dependent near-term outcomes.
Core views
China's domestic auto market has weakened materially. In the first eight months of 2026, wholesale volumes declined 5.3% year on year, with domestic demand down 19.7% and exports up 71.8%; August domestic SAAR was estimated at 19.4 million units, below 20.7-20.8 million in June and July. Industry participants attributed weaker demand to changes in subsidies, high fuel prices affecting ICE-heavy and joint-venture brands, depressed used-car residual values and dealer financial pressure. The China Association of Automobile Manufacturers expects second-half conditions to remain broadly similar to the first half, with no meaningful recovery in end-market demand. Nevertheless, NEV sales continue to grow, and participants expect EV penetration to reach 80-90% by 2030 as BEVs gain share. Exports are the principal offset to domestic weakness but bring policy and demand risks. CAAM cited Jan-August exports of 7.15 million units, up 66.7% year on year, with NEV exports rising more than 120% and representing nearly half of total exports. It estimates total 2026 exports could approach 11 million units, but warns that shipped vehicles may exceed actual overseas retail demand: it estimates a 1.5 million-unit gap in the first half between exports and overseas retail sales. Investors are focused on a potential reduction in China's 13% export VAT rebate. Companies view a phased reduction to 9% as more likely than immediate abolition, and BYD and Geely believe a 4 percentage-point reduction could largely be offset through cost optimization and supply-chain efficiency; a move to zero would likely require some customer pass-through. OEMs are also expanding overseas production to address potential EU PHEV tariffs and proposed 70% localization requirements. Competitive intensity and weak pricing are compressing profitability, supporting the case for consolidation. CAAM data show a 3.6% auto-manufacturing profit margin in Jan-Jul 2026 and a 20.4% year-on-year fall in industry profits; the association estimates many OEM margins are below 2%. Although headline prices have stabilized, competition has shifted into financing incentives, subsidies and faster product refreshes. The report sees government pressure against overcapacity, including possible GAC-FAW restructuring, as an early indication of more capacity rationalization. Industry experts expect companies seeking capacity increasingly to acquire or consolidate existing sites rather than build greenfield plants. Suppliers also reported better payment terms and more disciplined annual price negotiations after anti-involution initiatives. Vertical integration is becoming a core response to cost, supply-chain and technology pressure. Li Auto is broadening cell sourcing while integrating battery packs internally; Xiaomi is developing its Longjia battery platform and expanding into battery development, packaging and integration; OEMs including NIO, XPeng, Li Auto, Xiaomi and potentially BYD are advancing proprietary ADAS chips. Management teams argue that in-house capability improves supply security and hardware-software integration while lowering cost. XPeng said cumulative savings from replacing external compute solutions already exceed the R&D cost of its current Turing chip generation. The report notes that investors remain concerned about technology competitiveness against leaders such as Nvidia, but companies regard the economics as compelling. BYD's strategy centers on technology-led market-share gains, global expansion and an integrated flash-charging ecosystem. Management said domestic share reached roughly 25-26% without significant price cuts, supported by flash charging and next-generation batteries. Overseas targets of more than 1.8 million units in 2026 and 2.5 million in 2027 appear achievable, with a longer-term goal of 5 million overseas sales before 2030 excluding North America. Local capacity in Brazil, Indonesia and Hungary is intended to improve market access and mitigate tariffs. BYD argues that its advantage is the combined architecture of batteries, vehicles, charging infrastructure, energy storage and power electronics, which may be harder to replicate than standalone technology. Second-generation Blade Battery capacity is the near-term bottleneck, with output expected to increase by roughly 20,000 units per month; BYD shipped 60GWh of energy-storage systems in 2025 and expects improving mix, overseas contribution and utilization to support profitability. Other OEM discussions point to differentiated routes through the downturn. Xiaomi is targeting a 20% vehicle gross margin for SkyNomad, dependent on mix and accessory attachment, and plans its initial European launch in Germany in the second half of 2027. Geely sees exports as its principal growth engine, with overseas retail sales approaching 80,000 units per month and overseas profit per vehicle of RMB12,000-15,000; its asset-light localization strategy uses partners including Ford, Volvo, Renault and Proton rather than greenfield factories. XPeng expects overseas deliveries to at least double from roughly 24,000 units this year, product growth to reaccelerate in the fourth quarter after a flat third quarter, and stable second-half vehicle gross margins despite AI component inflation. Its delivery bottleneck at a low-value supplier nevertheless illustrates the importance of supplier qualification as launch cadence rises. NIO highlights a swap-station network of more than 4,000 sites and services income, but its average roughly 30 daily swaps per station remains below management's 50-60 swaps-per-day breakeven threshold. Li Auto expects intelligence, chips and AI models to become more important sources of differentiation than vehicle hardware alone, while continuing to view China NEV penetration and share gains from ICE and joint-venture brands as the main opportunity. Battery and energy-storage findings are selective. Solid-state batteries are still several years from large-scale deployment because of unresolved technical challenges, with early use more likely in humanoid robots and eVTOLs than passenger EVs. Sodium-ion adoption is viewed mainly as a function of lithium prices; BYD already has sodium-ion capacity and is using the technology in energy storage. Industry participants are more constructive than investors on energy storage, expecting 20-50% demand growth next year and longer-term CAGR of about 20% for China and 30% overseas through 2030. Autonomous-driving competition is moving from L2++ toward L3, L4 robotaxis and robovans. The report cautions that L3 differentiation could be constrained if technology commoditizes faster than expected. Data are widely viewed as the critical enabler, although participants differ over the relative importance of real-world and synthetic data; robotaxi deployment remains constrained chiefly by regulatory approval and licensing. Momenta cites a 60% share of China's urban NOA software market, 219 nominations across 26 OEMs, and 105 programs at SOP, with L3 projects under way and the Audi E7X expected as its first L3 production vehicle in the second half of 2027. Pony.ai reports robotaxi unit-economic breakeven in Guangzhou in November 2025 and Shenzhen in January 2026, a fleet of roughly 2,000 by mid-year targeted to reach 3,500 by year-end, and a Generation 7 BOM targeted below RMB240,000. Its view is that true L4 requires reinforcement learning, world models and synthetic data to safely address rare edge cases beyond L2+ imitation learning. Humanoid robotics is strategically important but not yet a broad earnings driver. Most OEMs and suppliers believe meaningful mass adoption remains three to five years away. XPeng has one of the faster schedules, with robot mass production targeted around early 2027 and early deployment in stores and showrooms, while BYD is evaluating manufacturing use cases where labor replacement and productivity gains could matter. The report also flags a U.S. humanoid customer's potential October production start and weekly capacity of about 2,000 units by year-end as a supply-chain inflection point, while emphasizing uncertainty over each supplier's sustainable share. Geopolitical exposure remains unresolved: companies point to manufacturing in Mexico and Thailand as mitigation, but investors question whether overseas production alone will circumvent future trade barriers.
Analysis framework
Bernstein combines management meetings, plant visits, vehicle test rides and discussions with industry experts and CAAM. It assesses demand, exports, policy, profitability, capacity, technology roadmaps, localization and supply-chain economics, then connects these industry observations to company-specific operating strategies, ratings, price targets and valuation comparisons.
Methodology notes
Auto demand, export volumes, capacity rationalization and NEV penetration
The report evaluates a weak domestic market against rapid exports, changing electrification penetration and excess capacity to explain the pressure on volumes and profitability.
OEM vertical integration and automotive supply-chain resilience
The report links OEM decisions on batteries, chips and overseas production to supplier economics, component availability, cost savings and trade-policy exposure.
Volume growth, pricing pressure, mix and margin progression
The analysis distinguishes vehicle volumes, export growth, pricing and specification upgrades, product mix and component costs to interpret profitability.
Reported P/E valuation comparisons
The report presents current and forward P/E comparisons for covered autos and robotics companies alongside earnings estimates and price targets.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- BYD (1211.HK / 002594.CH)Covered OEM; rated Outperform
- Strengths
- Technology-led domestic share gains, integrated flash-charging ecosystem, overseas expansion and energy-storage scale.
- Weaknesses
- Second-generation Blade Battery capacity is a near-term production bottleneck.
- Comparison
- Management argues its integrated ecosystem is harder to replicate than individual technologies.
- Risks
- Export VAT-rebate cuts, tariff and localization requirements, and the need for potential customer cost pass-through if rebates are eliminated.
- Geely (175.HK)Covered OEM; rated Outperform
- Strengths
- Exports, overseas profit per vehicle of RMB12,000-15,000, and asset-light localization partnerships.
- Weaknesses
- Domestic NEV growth is moderating as penetration matures.
- Comparison
- Uses existing partner capacity rather than primarily pursuing greenfield overseas factories.
- Risks
- Competitive domestic market, export-policy changes and execution of overseas localization.
- Xiaomi (1810.HK)Covered OEM; rated Outperform
- Strengths
- Core-chip development, Human × Car × Home ecosystem and expanding vehicle portfolio.
- Weaknesses
- SkyNomad margin depends on mix and accessory attachment rates.
- Comparison
- Management emphasizes control of both operating systems and hardware endpoints across devices, cars and homes.
- Risks
- Longer customer conversion cycle for SkyNomad and execution of European launch plans.
- XPeng (XPEV.US / 9868.HK)Covered OEM; rated Market-Perform
- Strengths
- Turing-chip cost savings, real-world data, product launches, overseas-margin potential and Physical AI roadmap.
- Weaknesses
- MONA deliveries were constrained by a component-supplier bottleneck.
- Comparison
- Management argues consumer-vehicle platforms provide robotaxi cost advantages over bespoke designs.
- Risks
- Supply-chain disruption, AI-component inflation, launch execution and accelerated autonomous-driving competition.
- NIO (NIO.US / 9866.HK)Covered OEM; rated Market-Perform
- Strengths
- More than 4,000 swap stations, services ecosystem, proprietary technology stack and expected volume growth.
- Weaknesses
- Average station utilization of about 30 daily swaps is below the 50-60 breakeven threshold.
- Comparison
- Management sees the battery-swap network effect as increasingly difficult to replicate.
- Risks
- Capital intensity, swap-network utilization and execution of growth in premium BEVs.
- Li Auto (LI.US / 2015.HK)Covered OEM; rated Market-Perform
- Strengths
- Vertical integration in batteries and chips, AI and ADAS investment, and product-refresh plans.
- Weaknesses
- Weak macro consumption and intensified premium-segment competition.
- Comparison
- Management expects future differentiation to shift toward chips, AI models and software rather than hardware alone.
- Risks
- Competitive pressure from Tesla, Huawei/AITO, Xiaomi and Zeekr, plus uncertainty around overseas execution.
- Hesai (HSA / 2525.HK)Covered emerging-robotics company; rated Outperform
- Tuopu (601689.CH)Covered emerging-robotics company; rated Outperform
- Shuanghuan (002472.CH)Covered emerging-robotics company; rated Outperform
- Sanhua (2050.HK / 002050.CH)Covered emerging-robotics company; rated Market-Perform
Key data
- China wholesale auto volumes-5.3% YoY in 2026 8MDomestic demand fell 19.7% while exports rose 71.8%.
- China auto domestic SAAR19.4 million units in August 2026Below 20.7-20.8 million units in June and July.
- China auto-industry profit margin3.6% in Jan-Jul 2026Total industry profits declined 20.4% YoY; CAAM estimates many OEM margins are below 2%.
- China Jan-Aug auto exports7.15 million unitsUp 66.7% YoY; NEV exports grew more than 120% YoY.
- Export-to-overseas-retail gap1.5 million units in 1H26CAAM estimate indicates exports may be ahead of final end-market sales.
- Long-term EV penetration80-90% by 2030Industry participant expectation.
- BYD overseas sales targets>1.8 million units in 2026; 2.5 million in 2027Management's targets; long-term goal is 5 million overseas sales before 2030 excluding North America.
- NIO swap-station utilizationc.30 swaps per dayBelow management's 50-60 daily swaps-per-station breakeven threshold.
- Momenta urban NOA market share60%Management estimate for China's urban NOA intelligent-driving software market.
- Pony.ai robotaxi fleet target3,500 vehicles by year-endFleet was roughly 2,000 by mid-year.
Impact & implications
Bernstein's tour findings suggest that leading OEMs may be better positioned to withstand a weak domestic cycle where exports, technology ownership, scale and localization can support resilience. However, the report presents robotics, robotaxis and higher-level autonomy as longer-duration opportunities whose value realization depends on commercialization, regulation, customer adoption and supplier execution.
Risks
- Domestic China auto demand may remain weak amid soft consumer spending, limited policy support and dealer financial pressure.
- Export growth may exceed final overseas demand, as evidenced by CAAM's estimated 1.5 million-unit first-half gap between exports and overseas retail sales.
- A reduction or removal of China's export VAT rebate could pressure OEM economics and potentially require price pass-through.
- Future tariffs and localization rules may not be fully mitigated by overseas manufacturing footprints.
- Intense competition, indirect pricing incentives and excess capacity may continue to pressure industry profitability.
- Autonomous-driving differentiation could be limited if L3 technology commoditizes faster than expected.
- Robotaxi rollout remains subject to regulatory approvals and licensing.
- Humanoid-robot commercialization and individual suppliers' sustainable market shares remain uncertain.
What to watch
- China domestic auto demand, subsidy policy and the pace of industry consolidation.
- Changes to China's 13% export VAT rebate, EU PHEV tariffs and localization requirements.
- Whether export shipments translate into overseas retail demand.
- OEM progress in battery, chip and ADAS vertical integration and associated cost savings.
- BYD battery-capacity expansion, international flash-charging rollout and overseas sales delivery.
- XPeng's product-launch execution and supplier qualification following the MONA bottleneck.
- NIO swap-station utilization relative to the 50-60 swaps-per-day breakeven threshold.
- L3 approvals, robotaxi licensing and the commercial scaling of humanoid robots.