Two weeks ago I wrote about my 2026 investment strategy – targeting 15-25% returns in Consumer Discretionary, Industrials, and Health Care. Each week I look at one stock from these sectors.
Last week was ATI, a US specialty metals supplier riding defence spending but trading at stretched valuations. Gold and silver crashed on 30 January, but ATI’s stock held steady – showing that government defence contracts matter more than raw metal prices for this company.
This week: XPeng versus Tesla – same technology roadmap, but both stocks are priced very differently. Both are building EVs, autonomous driving, and humanoid robots. One is growing 150%, the other declining 8.6%. Here’s why I’m backing the cheaper one.
The Company
XPeng is building a full-spectrum mobility company – EVs, autonomous driving, humanoid robots, and flying cars. They’re executing the same product roadmap as Tesla but in China at mass-market prices.
Founded in 2014, XPeng delivered 429,000 vehicles in 2025 (+126% YoY), capturing roughly 4% of China’s EV market. That makes them the 6th largest EV maker in the world’s largest EV market – smaller than BYD (4.3 million units) but ahead of established players like NIO and Li Auto.
What they’re building:
| Category | Tesla | XPeng |
|---|---|---|
| Electric Vehicles | Model S/3/X/Y/Cybertruck | P7+, G6, G9, X9, MONA M03 |
| Autonomous Driving | FSD – Level 2 (driver supervises) | XNGP / VLA 2.0 – vision-only (Q1 2026 launch) |
| Robotaxis | Austin pilot with safety drivers | Three models launching 2026 |
| Humanoid Robots | Optimus (demo stage) | IRON (prototype complete, mass production end 2026) |
| Flying Cars (eVTOL) | None | Land Aircraft Carrier – deliveries late 2026 |
| Charging Network | Supercharger network | 2,676 stations globally (as of Sept 2025) |
| Energy Products | Powerwall, Megapack | None – pure mobility play |
The comparison matters because both companies are pursuing identical strategies at vastly different valuations.
Why I’m Invested
XPeng trades at 2x sales whilst approaching profitability. Tesla trades at 16x sales and 300x earnings with declining deliveries.
Both are building the same future. One is executing, the other is promising.
The operational divergence:
Q3 2025 results tell the story:
| Metric | XPeng | Tesla |
|---|---|---|
| Revenue Growth | +102% YoY | -3% (first annual decline) |
| Delivery Growth | +149% (116K units) | -8.6% for full year |
| Gross Margin | 20.1% (first time above 20%) | Declining |
| Bottom Line | Lost US$50M (approaching breakeven) | US$840M profit (down 61% YoY) |
| 2026 Guidance | 550K-600K vehicles (+28-40%) | No guidance provided |
Note: Using Q3 2025 quarterly data as XPeng’s Q4 2025 results release on March 24, 2026 – after this article’s February 6 publication date.
XPeng just achieved 20% gross margins via their SEPA 2.0 platform – modular architecture cutting costs 25-50% per component. Tesla’s margins are compressing due to price cuts and an aging product lineup.
XPeng delivered 429,000 vehicles in 2025 (+126%). Tesla delivered 1.64 million (-8.6%), recording its first annual revenue decline. Tesla also announced it will end production of its Model S sedan and Model X SUV in Q2 2026 – discontinuing its two flagship vehicles that have been in production since 2012 and 2015. The factory space will be converted to produce Optimus robots instead. This leaves Tesla with just three vehicles: Model 3, Model Y, and Cybertruck.
Where XPeng makes money:
XPeng’s business is concentrated in China (~90% of deliveries), with growing international presence (~10% of deliveries in 2025, up 92% YoY). As of December 2025, the company operates 1,101 stores across 60 countries including:
- Europe: 28 countries, 290 stores, local production in Austria
- Southeast Asia: Malaysia, Thailand, Indonesia with local assembly
- Middle East: Premium positioning, Egypt service centre
90% China concentration is both a risk and an advantage – they compete in the world’s largest EV market (30.1M passenger vehicle sales, 16.5M NEV sales in 2025) while avoiding US tariff exposure entirely.
The autonomous driving gap:
Tesla promised Full Self-Driving in 2016. Eight years later, it’s still Level 2 – driver must supervise at all times.
XPeng’s VLA 2.0 is launching March 2026. The technology has been validated through academic publication (accepted at AAAI 2026, top AI conference) and commercial partnership – Volkswagen is paying US$700M to license XPeng’s autonomous driving stack.
When Europe’s largest automaker partners with a Chinese startup rather than developing internally, that signals where the technological edge exists.
The valuation makes sense:
| Criteria | Assessment | Details |
|---|---|---|
| Financial Health | ✓ Strong trajectory | US$6.8B cash vs US$2.4B debt (net cash positive). Debt-to-equity 43% and improving. Lost only US$50M in Q3 2025, targeting breakeven Q4. |
| Growth Drivers | ✓ Multiple engines | +149% deliveries, international expansion (+92% overseas), VW licensing revenue, full-spectrum mobility (EVs + autonomy + robotics + flying cars). |
| Valuation | ✓ Attractive | ~2x sales vs analyst targets US$27-29 (47-56% upside). Tesla at 16x sales, 300x earnings for comparison. |
| Risk Management | What could go wrong | China concentration (90% revenue), brutal competition, profitability not guaranteed, autonomous driving execution uncertain. |
The Numbers (Q3 2025)
Balance sheet: Net cash positive with US$6.8 billion in cash versus US$2.4 billion in debt. Debt-to-equity ratio of 43%, down from 57.7% five years ago – improving trend.
Profitability path: Q3 2025 net loss was US$50 million, down from US$1.8 billion a year earlier. Management is guiding to breakeven in Q4 2025.
The operational turnaround is real. Three factors drove it:
- Volume – Selling 4x more cars spreads factory costs
- Product-market fit – The US$17K MONA M03 and P7+ are achieving strong sales momentum
- Smart engineering – SEPA 2.0 platform cuts costs 25-50%
Cash generation: Not yet free cash flow positive, but the trajectory is clear. At 20% gross margins and approaching 500K+ annual volume, the unit economics work.
Regulatory tailwind: In December 2025, China’s market regulator drafted guidelines to halt below-cost pricing after years of destructive price wars. The move targets “irrational competition” that pushed industry profit margins to 3.9% (down from 7.8% in 2017) and forced 52% of dealers into losses. XPeng publicly supported the regulations. If enforced, this shifts competition from price to technology and quality – precisely where XPeng’s autonomous driving IP and VW licensing revenue give them an edge. The timing benefits XPeng at 20% gross margins while weaker rivals struggle near breakeven.
Tesla’s problem is the opposite. They’re profitable but making less money despite higher volume. Competition is eating their lunch in China and Europe.
Why I Think They’ll Keep Growing
1. China market momentum
China is the world’s largest EV market – 30 million annual vehicle sales, over 10 million EVs in 2025. NEVs (new energy vehicles) outsold traditional ICE vehicles for the first time in 2025.
XPeng delivered 429,000 vehicles in 2025, up 126%. Tesla’s China sales declined.
The Chinese consumer is adopting smart car technology faster than Western markets. Autonomous driving features, in-car AI assistants, and OTA updates are expected features, not premium add-ons.
2. International expansion
Overseas deliveries hit 45,000 units in 2025 (+92% YoY), representing 10.5% of total volume. 2026 target: double overseas sales to ~90,000 units.
Key markets:
- Europe: 28 countries, 290 stores, local production via Magna in Austria (approximately half of overseas sales)
- Middle East & Africa: Largest service centre in Egypt
- Asia-Pacific: Malaysia, Thailand, Indonesia with local CKD assembly
Why international matters: European and ASEAN markets command 15-20% higher average selling prices versus China whilst maintaining similar cost structures. This margin enhancement is material as international mix grows.
CEO target: “Half of XPeng’s sales will come from global markets by 2035.” From a 550K-600K 2026 base, that implies 1.1-1.2 million units by 2035, with 550K-600K international.
3. Technology licensing
Volkswagen is paying US$700M to license XPeng’s autonomous driving technology and Turing AI chips. Revenue started Q4 2025.
Breaking: VW began production using XPeng’s CEA architecture (announced Jan 29, 2026). The first VW-branded vehicle using XPeng’s jointly-developed electronic architecture entered production at VW’s Anhui plant in late 2025. Five additional VW models will launch in 2026 using the CEA platform, covering EVs, hybrids, and combustion engines across VW’s three China joint ventures.
VW reports the CEA architecture delivers 30% higher development efficiency and cuts costs by 50% for key models. The partnership has already generated ¥1.44 billion (~US$200M) in technical services revenue for XPeng in Q1 2025 alone.
This validates two things: (1) XPeng’s technology is commercially viable and in actual production, and (2) they’ve built a licensing revenue stream beyond vehicle sales.
4. Full-spectrum mobility strategy
Like Tesla, XPeng isn’t just an EV manufacturer:
- Core business: Electric vehicles (429K units in 2025)
- Autonomous driving: VLA 2.0 vision-only system launching March 2026
- Robotics: IRON humanoid robot, first automotive-grade prototype completed January 2026
- Flying cars: Controlling stake in AeroHT, developing eVTOL for urban mobility
This creates multiple paths to value creation beyond traditional automaking. The AI and autonomy platform applies across mobility categories.
5. No US tariff exposure
Critical nuance: XPeng currently doesn’t sell in the United States. Zero exposure to 125% combined tariffs (100% EV-specific + 25% base rate) that make Chinese vehicles uncompetitive in the US market.
This is actually strategic protection. XPeng can win big focusing on China + Europe + Asia, which represent 60% of the global EV market, without touching the geopolitically fraught US market.
Tesla, by contrast, derives 45% of revenue from the US – directly in the line of potential tariff retaliation.
6. Autonomous driving execution
Tesla promised Full Self-Driving in 2016. Eight years later, it’s still Level 2 – driver must supervise at all times.
XPeng’s VLA 2.0 is launching March 2026. The technology has been validated through academic publication (accepted at AAAI 2026, top AI conference) and commercial partnership – Volkswagen is paying US$700M to license XPeng’s autonomous driving stack.
When Europe’s largest automaker partners with a Chinese startup rather than developing internally, that signals where the technological edge exists.
Timeline comparison:
- Tesla: 2016 announcement → 2025 still Level 2 → No production timeline for robotaxis
- XPeng: 2023 XNGP launched → 2024 expanded to 243 cities → March 2026 VLA 2.0 launches
Both companies use vision-only systems (cameras, no LiDAR). Waymo’s LiDAR-dependent system costs approximately US$200,000 per vehicle. XPeng’s vision-only system operates in US$17,000-30,000 consumer vehicles.
If vision-only succeeds, it scales to billions of vehicles. The question isn’t which technology is theoretically better – it’s which company is actually shipping.
7. Robotics upside
Tesla announced Optimus in 2021. Production timeline keeps slipping – initial plans for thousands of units by 2025 pushed to 2026. October 2024 demos were criticized when attendees discovered robots were remotely controlled by humans.
XPeng unveiled IRON in November 2025. January 19, 2026: Completed first ET1 robot prototype developed to automotive manufacturing standards. Mass production planned end of 2026.
Neither has proven commercial viability – this represents upside optionality, not current value. But the pattern of execution versus promise matters.
8. Flying cars – another optionality play
XPeng’s AeroHT subsidiary (now branded Aridge) began trial production of the “Land Aircraft Carrier” in November 2025 at a dedicated 120,000 sq meter factory in Guangzhou. This is the world’s first flying car mass production line using automotive-style assembly.
The product is modular: a 6-wheel ground vehicle (5.5m long, drivable with standard license) with a detachable 2-seater eVTOL aircraft that stores in the trunk. The aircraft can fly while the mothership charges it between flights.
Target: Mass deliveries late 2026 at under ¥2 million (US$280K). Already secured ~7,000 orders (90% domestic, 10% overseas).
Tesla has no flying car program. XPeng is building the infrastructure and manufacturing capacity for what China calls the “low-altitude economy” – a government-designated strategic sector with 1,200+ takeoff/landing sites already under construction in cities like Shenzhen.
What Am I Paying For?
Current valuation:
- Stock price: US$16.88 (NYSE, Feb 5, 2026 close) / HK$67.30 (HKEX)*
- Market cap: US$19 billion
- Price-to-sales: ~2x (approaching profitability)
- Wall Street consensus: US$25-28 (NYSE), representing 48-66% upside
*Note: 2 NYSE ADRs ≈ 1 HKEX share.
What the market is pricing:
At 2x sales, the market is valuing XPeng as a pure EV player, not a mobility company. The current price reflects significant execution uncertainty around achieving profitability and sustaining growth. The autonomous driving stack (validated by VW’s $700M investment and technology partnership), robotaxis, flying cars (7,000 orders), and humanoid robots are essentially being given zero value.
Wall Street’s $27-29 consensus reflects successful execution on the core EV business – achieving profitability and sustaining delivery growth. But even these analyst targets don’t price in the mobility categories. If XPeng executes on robotaxis, flying cars, or humanoid robots, the valuation could re-rate significantly beyond current Street estimates.
What Could Go Wrong?
Here’s what I’m watching:
1. China concentration with strategic upside
90% of revenue comes from China. Any domestic slowdown or regulatory change hits directly.
However, this concentration has an upside: XPeng avoids US-China geopolitical risk entirely. The company doesn’t sell in the United States and is immune to 125% combined tariffs.
XPeng’s international expansion focuses on Europe (28 countries, 290 stores) and Southeast Asia (Malaysia, Thailand, Indonesia with local assembly). China + Europe + Asia represent 60% of the global EV market – XPeng can win without touching US soil.
2. Competition intensity in China market
The China EV market is structurally oversupplied. BYD (28% market share), Li Auto, NIO, plus 100+ manufacturers fighting for share in a market with severe overcapacity – battery industry utilization is only 41%.
The financial stress is visible everywhere:
- Industry profit margins collapsed from 7.8% (2017) to 3.9% (Q1 2025)
- 52% of dealers operated at losses in H1 2025
- Automakers taking 108 days on average to pay suppliers (up from 99 days in 2019) – a sign of cash flow pressure
- BYD taking 127 days to pay suppliers despite being market leader
Many competitors are highly leveraged and losing money at scale. The market consolidated from 500 manufacturers in 2019 to roughly 100 in 2025, with AlixPartners forecasting that only ~15 brands will remain financially viable by 2030.
XPeng is growing 150% whilst competitors like Li Auto are contracting (deliveries down 13% YTD 2025, falling four consecutive months). But if industry consolidation triggers a deeper price war before regulations take hold, even strong players face margin pressure.
3. Profitability uncertainty – almost there but not proven
XPeng is approaching breakeven but hasn’t crossed the line yet. Q4 2025 target (earnings release March 24, 2026) is management guidance, not guaranteed result.
The trajectory is encouraging – from losing US$1.8 billion a year ago to just US$50 million last quarter. At 20% gross margins and 116K quarterly deliveries, the unit economics work mathematically.
But “almost profitable” means execution risk is magnified. If Q4 2026 volumes disappoint due to competition, or if the price war intensifies before regulations take hold, losses could persist into 2026. The company needs to maintain 20%+ gross margins and 120K+ quarterly volume to reach profitability. Any slip in either metric pushes breakeven further out.
4. Autonomous driving execution
Consumer AV adoption may take longer than expected. Vision-only systems may prove insufficient for Level 4 autonomy, requiring expensive hardware retrofits. Regulatory approval remains uncertain across markets.
The March 2026 VLA 2.0 launch is a milestone, but commercial success isn’t guaranteed.
5. Robotics remains speculative
IRON robot mass production is planned for end of 2026, but commercial viability is unproven. This represents upside optionality, not current value. If humanoid robots don’t achieve product-market fit, this thesis component disappears.
Conclusion
2026 will be a transitional year for XPeng.
Five milestones will define the story: achieving breakeven (Q4 2025 results in March), launching VLA 2.0 autonomous driving (Q1), deploying robotaxis (2026), delivering flying cars (late 2026), and starting IRON robot mass production (end of year).
If they execute on these, XPeng shifts from “promising EV startup” to “profitable autonomous mobility company.” That’s the re-rating I’m betting on.
This article represents my personal opinion and investment decisions. It is not financial advice or investment research. I hold positions in securities discussed. All investors should conduct independent due diligence and consult qualified financial advisors before making investment decisions. XPeng trades on NYSE (XPEV) and HKEX (9868.HK).