In-Depth Evaluation Report on XPeng Motors (XPEV) Strategic Transformation
Unlock More Features
Login to access AI-powered analysis, deep research reports and more advanced features

About us: Ginlix AI is the AI Investment Copilot powered by real data, bridging advanced AI with professional financial databases to provide verifiable, truth-based answers. Please use the chat box below to ask any financial question.
Related Stocks
Based on the above in-depth research, I provide you with a systematic analysis report on XPeng Motors’ strategic transformation.
XPeng Motors is undergoing the most profound strategic transformation since its establishment. At the global new product launch in January 2026, He Xiaopeng clearly put forward the new positioning: ‘In the last decade, XPeng was representative of intelligent electric vehicles; in the new decade, XPeng is representative of Physical AI and globalization’ [1]. This marks that the company has shifted from pure product competition to technological ecological competition.
| Pillar | Core Content | Implementation Progress |
|---|---|---|
| Self-developed Chips | Turing AI Chip, mass-produced and installed on vehicles | Tape-out successful in August 2024, mass production started in Q2 2025 |
| Algorithm Iteration | 2nd-generation VLA Large Model, end-to-end intelligent driving | Robotaxi has passed third-party site testing |
| Ecological Expansion | Humanoid robots, flying cars, Robotaxi | Humanoid robot IRON has been unveiled |
According to the research report from Soochow Securities, XPeng has built a complete technology triangle of ‘self-developed large model (algorithm) + self-developed Turing chip (computing power) + tens of millions of Clips data (data)’ [2]. This vertical integration model is highly similar to Tesla’s path, but XPeng has chosen a more open ecological strategy — in November 2025, XPeng announced that its second-generation VLA intelligent driving system and Turing chip had been designated by Volkswagen Group, becoming its first strategic partner [3].
He Xiaopeng predicts that in the next decade, the value of automotive hardware will drop from 90% to 50%, and software will increase from 10% to 50% [1]. The core logic of this judgment lies in:
- Hardware Standardization: With the increase in electrification penetration, the three-electric system (battery, motor, electronic control) is gradually homogenized
- Software Differentiation: Intelligent driving and cockpit experience have become core competitive barriers
- Computing Power Sovereignty: He Xiaopeng believes that ‘the world’s best AI companies in the future will all self-develop chips’, which is not only a cost issue but also a necessary condition for achieving generational technological leaps [1]
- Vehicle gross margin reached 14.3% in Q2 2025, improving for 8 consecutive quarters [4]
- Global deliveries reached 429,000 units in 2025, a year-on-year increase of 126%, ranking second among new carmakers, surpassing Li Auto and Xiaomi Auto [3]
- Entered 60 countries and regions in 2025, targeting a 1:1 sales ratio between overseas and Chinese markets in 2026 [1]
According to calculations from Soochow Securities [2], self-developed chips have obvious cost advantages compared to externally procured NVIDIA Orin-X:
| Cost Item | External Procurement Solution (NVIDIA Orin-X) | Self-Developed Solution (Turing) |
|---|---|---|
| Chip Price | ≥ RMB 2,800 | - |
| Development Cost | - | RMB 200 |
| Manufacturing Cost | - | RMB 800 |
| Packaging and Testing Cost | - | RMB 100 |
| Labor Cost | - | RMB 500 |
Total |
RMB 2,800 |
RMB 1,600 |
Savings |
- | RMB 1,200 per chip |
- Annual production and sales of 300,000 vehicles, 2 chips per vehicle
- Annual cost savings: 300,000 × 2 × 1,200 = RMB 720 million
- Break-even point: Annual sales of approximately 130,000-150,000 vehicles can cover R&D costs
- However, it should be noted that the initial R&D investment for self-developed chips is approximately RMB 2-3 billion (7nm process), with a long payback period
However, industry insiders point out that it is difficult to recoup the cost of self-developed chips in the short term. A single tape-out requires tens of millions of US dollars in preparation, and if switching to a 5nm process, it is necessary to queue at TSMC, and a large amount of reserves are required [3]. XPeng’s solution is to amortize R&D costs through external sales — the designated cooperation with Volkswagen Group and negotiations with Hyundai Motor are clear evidence.
He Xiaopeng claimed that ‘one self-developed Turing chip is equivalent to three traditional chips’, and its performance advantages are reflected in:
| Indicator | XPeng Turing | NIO Shenji NX9031 | NVIDIA Orin-X |
|---|---|---|---|
| Launch Time | August 2024 | July 2024 | 2019 |
| Computing Power | ~750TOPS | ~1000TOPS | 254TOPS |
| Process | 7nm | 5nm | 7nm |
| CPU | 40 cores | 32 cores | 12 cores |
| ISP | 2 independent units | High-performance | 1 unit |
More importantly,
- Eliminates the “language translation” link, enabling end-to-end generation from visual signals to control commands
- Improved computing power effectiveness — He Xiaopeng stated that 2250 TOPS of computing power from the Turing chip can achieve the actual performance of 4500 TOPS from competitors
- More flexible iterative upgrades, with collaborative optimization of algorithms and chips
According to test videos, vehicles equipped with the second-generation VLA can already handle:
- Temporary road construction and lane occupation, autonomous planning of lane borrowing for passage
- Unprotected left turns, defensive deceleration and game-based passage
- Understanding gesture instructions from road construction workers [1]
Self-developed chips are not only a cost consideration but also strategic security:
- Supply Disruption Risk: Under geopolitical pressure, mastering core computing power is essential
- Technological Iteration: He Xiaopeng believes that ‘based on self-developed chips and compilation optimization, generational technological leaps can be achieved’ [1]
- Open Ecology: XPeng plans to package its intelligent driving, chip, and Robotaxi capabilities to attract external customers, forming diversified revenue sources

| Indicator | Q1 2024 | Q2 2024 | Q3 2024 | Q4 2024 | Q1 2025 | Q2 2025 | Q3 2025 |
|---|---|---|---|---|---|---|---|
| Vehicle Gross Margin | 5.5% | 6.4% | 8.3% | 9.5% | 10.5% | 14.3% | 14.0% |
| Comprehensive Gross Margin | 12.9% | 13.5% | 14.2% | 15.1% | 15.6% | 17.3% | 16.8% |
| Deliveries (thousands of units) | 21.8 | 23.2 | 30.5 | 35.4 | 47.4 | 51.6 | 46.0 |
| Revenue (billion RMB) | 6.5 | 8.1 | 10.1 | 13.4 | 15.8 | 17.2 | 16.5 |
- Gross margin improved for 8 consecutive quarters: Increased from 5.5% in Q1 2024 to a peak of 14.3% in Q2 2025, with an improvement of 8.8 percentage points
- Economies of scale emerged: Deliveries grew from 21,800 units to 51,600 units (+137%), significantly amortizing unit fixed costs
- Product structure optimization: Increased proportion of high-margin models G6, G9, and P7+ drove the overall gross margin upward
Based on the current trend and company guidance, achieving this goal requires:
| Condition | Current Status | Target Value |
|---|---|---|
| Comprehensive Gross Margin | 17.3% (Q2) | ≥20% |
| Vehicle Gross Margin | 14.3% (Q2) | ≥18% |
| Quarterly Deliveries | 51,600 units (Q2) | ≥60,000 units |
| Revenue | RMB 17.2 billion (Q2) | ≥RMB 20 billion |
- Actual EPS of -$0.06 in Q3 2026, lower than the expected $0.02 (-371.62%)
- Revenue of $2.86 billion, lower than the expected $3.30 billion (-13.26%)
- R&D expenses in 2025 are approximately RMB 10 billion (a 54% increase from RMB 6.5 billion in 2024) [3]
- The Chinese passenger vehicle market faces downward pressure in 2026 (preferential purchase tax policy withdrawal) [3]
- Clear trend of gross margin improvement
- Sustained growth in deliveries
- Sufficient cash reserves (approximately RMB 33.63 billion as of the end of March 2025)
- Analysts maintain a Buy rating, with a target price of $29.50 (+47.4% from current price) [4]
| Risk Category | Risk Description | Probability | Impact Level |
|---|---|---|---|
Execution Risk |
Long payback period for self-developed chips, economies of scale fall short of expectations | Medium | High |
Market Risk |
Persistent price war, pressure on gross margin improvement | High | Medium |
Technological Risk |
Delayed landing of L4-level intelligent driving, Robotaxi commercialization falls short of expectations | Medium | Medium |
Capital Risk |
Continuous increase in R&D investment, cash flow pressure increases | Medium | High |
Competitive Risk |
Improved competitiveness of intelligent driving solutions from competitors such as Huawei and Momenta | High | Medium |
| Automaker | Chip | Process | Computing Power | Positioning |
|---|---|---|---|---|
| XPeng | Turing | 7nm | ~750TOPS | Cost-effective priority, open ecology |
| NIO | Shenji NX9031 | 5nm | ~1000TOPS | Performance priority, closed ecology |
| Li Auto | M100 | - | - | Deliveries start in 2026 |
- Huawei: Smart Selection model provides brand output, regarded as the “Apple of intelligent driving solutions”
- Momenta: Fast delivery capability, regarded as the “Android of intelligent driving solutions”
- NVIDIA: Orin-X still dominates, with continuous iteration of the Thor series
XPeng’s differentiation lies in:
- Most Open: Packages intelligent driving, chip, and Robotaxi capabilities for external output
- Earliest Landing: Turing chip has been mass-produced and installed on vehicles, with Volkswagen’s designation as endorsement
- Widest Ecology: Covers automobiles, robots, and flying cars
- ✓ Clear trend of sustained gross margin improvement (+2 points)
- ✓ Self-developed chip has been mass-produced and received external designation (+2 points)
- ✓ Rapid growth in deliveries, economies of scale emerged (+2 points)
- ✓ Sufficient cash reserves, sustainable R&D investment (+1.5 points)
- ✓ Smooth progress of globalization strategy (+1 point)
- ✗ Still in loss, pressure from profitability commitment (-1.5 points)
- ✗ Persistent price war, industry under pressure (-1.5 points)
- ✗ Long payback period for self-developed chips (-1 point)
- ✗ Uncertainty in technological implementation (-1 point)
| Dimension | Evaluation |
|---|---|
Valuation Rationality |
Current P/S ratio of 1.94x, in the middle of the historical valuation range |
Upside Potential |
Analyst target price of $29.50, +47.4% from current price |
Risk-Return Profile |
Medium-high risk, medium-high return, suitable for risk-preferred investors |
Core Focus Areas |
1. Whether the profitability commitment can be fulfilled in Q4 2025 2. Whether gross margin can break through the 20% mark with sustained improvement 3. Commercialization progress of Robotaxi 4. Increase in overseas sales proportion |
XPeng Motors’ strategic transformation from hardware manufacturing to an AI company has a
- Cost Dimension: Saves approximately RMB 1,200 per chip, but requires economies of scale support (annual production and sales of over 300,000 vehicles for significant effect)
- Efficiency Dimension: Software-hardware integration improves computing power effectiveness, with 2250 TOPS achieving the effect of 4500 TOPS from competitors
- Strategic Dimension: Core support for supply chain security, technological sovereignty, and open ecology
- ✓ Whether the profitability commitment can be fulfilled in Q4 2025
- ✓ Whether the 1:1 sales target between overseas and Chinese markets can be achieved in 2026
- ✓ Whether Robotaxi can be commercialized as scheduled
- ✓ Whether gross margin can continue to improve and break through the 20% threshold
From the capital market performance (+69.52% in the past year), the market has relatively positive expectations for the transformation [4]. However, investors should note that XPeng is currently in a critical stage of “betting on the present with extended-range vehicles, and betting on the future with AI” [1]. The balance between short-term performance fluctuations and long-term strategic investment will be the core determinant of success or failure.
[1] Gasgoo - “In the second half of intelligent electric vehicles, XPeng’s gameplay has changed” (https://auto.gasgoo.com/news/202601/9I70441859C109.shtml)
[2] Soochow Securities - “Analysis of XPeng Motors’ Core Competitiveness” (https://pdf.dfcfw.com/pdf/H301_AP202504201659383369_1.pdf)
[3] The Paper - “The end of XPeng’s Physical AI is Musk’s cash flow” (https://m.thepaper.cn/newsDetail_forward_31973239)
[4] Gilin API Data - XPEV Real-time Quotation and Company Profile
[5] BeyondSXP - “XPeng’s AI Ascent: Powering Growth and Profitability” (https://beyondspx.com/quote/XPNGF/analysis/xpeng-s-ai-ascent-powering-growth-and-profitability-through-smart-ev-innovation-xpngf)
[6] Xchuxing - “Tripartite Competition of Self-Developed Chips, New Battlefield for Leading Intelligent Driving” (https://www.xchuxing.com/article/135742)
[7] Securities Times - “He Xiaopeng adjusts strategy again: betting on the present with extended-range vehicles, betting on the future with AI” (https://www.stcn.com/article/detail/3586381.html)
[8] Southern Plus - “Leading with AI, Four Cars Launched! What big moves does XPeng Motors have in the new year?” (https://www.nfnews.com/content/xovKnGOMyJ.html)
Insights are generated using AI models and historical data for informational purposes only. They do not constitute investment advice or recommendations. Past performance is not indicative of future results.
About us: Ginlix AI is the AI Investment Copilot powered by real data, bridging advanced AI with professional financial databases to provide verifiable, truth-based answers. Please use the chat box below to ask any financial question.
