Beijing time, September 2nd evening - NIO Inc. (NYSE: NIO; HKEX: 9866; SGX: NIO) released its second quarter 2025 financial results ending June 30th today: Total revenue reached RMB 19.009 billion, up 9.0% year-over-year and 57.9% quarter-over-quarter. Net loss was RMB 4.995 billion, narrowing by 1.0% year-over-year and 26.0% quarter-over-quarter. Under non-GAAP accounting standards, adjusted net loss was RMB 4.127 billion, narrowing by 9.0% year-over-year and 34.3% quarter-over-quarter.

Following the earnings release, NIO Inc. founder, chairman and CEO William Li, and CFO Stanley Qu participated in a conference call to discuss the results and answer analyst questions.

Below are excerpts from the analyst Q&A session:

**Citi Securities Analyst Jeff Chung**: My first question concerns the production capacity and ramp-up timeline for ES8 and L90, as well as delivery targets for the remainder of the year. Given the current severe order backlog, can we expect the group's single-month deliveries to reach 55,000 units or more in December?

**William Li**: The overall demand for L90 and ES8 is indeed very strong, stronger than we originally anticipated. We are working with supply chain partners to comprehensively boost production capacity. Our current plan targets L90 to achieve full supply chain capacity of 15,000 units in October, while for ES8, which has a capacity ramp-up process, we aim to reach 15,000 units of monthly full supply chain capacity by December. Based on this year's demand and supply situation, our current target is to achieve an average monthly delivery of 50,000 units in Q4, meaning over 150,000 deliveries for the fourth quarter.

**Jeff Chung**: My second question concerns gross margins and whether we can achieve break-even at the net profit level in Q4. We saw Q2 revenue grow 58% year-over-year, but gross profit increased over 100% quarter-over-quarter. Can you elaborate on vehicle gross margin trends for the second half and non-vehicle business gross margin trends? Specifically, how does management view the respective gross margin levels for L90 and ES8?

**Stanley Qu**: Let me explain the company's gross margin trends starting from Q2. Q2 vehicle gross margin was 10.3%. The main reason was that Q2 was a product transition period. Our 2025 "5566" models (ET5, ET5T, ES6, EC6) completed their transition in mid-to-late May, and from an overall sales mix perspective, among the 72,000+ Q2 deliveries, 5566 models only accounted for about 20% of deliveries. Therefore, Q2 vehicle gross margin improvement wasn't as significant compared to Q1.

In Q3, with the completion of NIO Inc. brand's 5566 transition and L90 starting deliveries, our vehicle gross profit will gradually improve. In Q4, as Li just mentioned, ES8 will begin deliveries from late September, and both L90 and ES8 will have full-quarter deliveries in Q4, so our gross margin target will improve further. When we achieve break-even in Q4, our vehicle gross margin target is between 16-17%.

Our ten years of pure electric technology accumulation, key component self-research, and continuous procurement and cost control measures enable us to not only create competitive products like L90 and ES8 but also achieve systematic comprehensive cost reduction. For Q4, our gross margin target for L90 and ES8 is 20%.

Other business profit margins achieved 8.2% positive gross margin in Q2 this year. This mainly benefited from our existing user base driving profit improvements in after-sales and financial services, narrowed losses in energy services, and importantly, we also achieved significant profits from external technical services in Q2. Overall, Q2 achieved substantial improvement compared to previous quarters.

Regarding the external technical services I mentioned, they will fluctuate with product cycle changes, so they're not stable every quarter. For other business gross margins, we hope to achieve break-even or maintain slight losses quarter-over-quarter.

**Credit Suisse Analyst Bin Wang**: I'd like to understand more details about the company achieving break-even in Q4. First, what are the R&D expenses for Q3 and Q4 respectively? I recall management previously mentioning Q4 R&D expense target approaching RMB 2 billion - is this guidance still maintained? Second, is the sales, general & administrative (SG&A) expense guidance also maintained? Finally, what specifically does "break-even" refer to? Is it operating profit (OP) level break-even or gross margin level break-even? Is it based on GAAP or non-GAAP?

**Stanley Qu**: Regarding the break-even target, we maintain non-GAAP break-even. For your two questions, regarding R&D guidance, we've actually done significant work on R&D expenses starting from Q2. Combined with our basic operating unit mechanism, the overall goal is to gradually improve R&D efficiency without affecting mainline output, meaning our upcoming product planning and R&D won't be affected. Second, we need to push the entire team to improve efficiency, so for overall R&D guidance, we still hope to achieve the quarterly non-GAAP RMB 2 billion target in Q3 and Q4.

Regarding SG&A, the company will combine the basic operating unit mechanism to gradually improve efficiency in this area. From Q2 perspective, with overall sales at 70,000 units, SG&A as a percentage of sales is still relatively high, but as sales volume and revenue increase, efficiency will gradually improve. In Q3, we have many new models launching, so overall marketing activities and expenses will be higher, and we can't achieve overall break-even yet. But by Q4, we hope to achieve non-GAAP SG&A as a percentage of sales below 10% to help us achieve Q4 non-GAAP break-even.

**Morgan Stanley Analyst Tim Hsiao**: I have two questions. First, regarding new model planning. Given strong demand for L90 and ES8 but capacity constraints, will the company adjust subsequent model release timelines? We also notice some new model launches have been significantly delayed to late September. Can management provide updates on new model planning for the coming quarters?

**William Li**: Indeed, from a capacity perspective, we're definitely prioritizing L90 and the new ES8. In the ONVO business, even L60 capacity needs to give way to L90, and ONVO L60 also needs to wait. So currently, our company has four models with order backlogs waiting for production and delivery: L90, new ES8, L60, and Firefly. The ONVO situation will see significant relief in October, mainly because battery capacity should reach reasonable levels starting in October - we've increased supply with battery partners over the past few months. Overall, we hope to increase ONVO's full supply chain capacity to 25,000 units monthly in Q4.

For the NIO Inc. brand, the main challenge is also the 102kWh battery used by the new ES8, whose demand is indeed better than our original plan, so our initial battery supply estimates were conservative. Currently, we're working with battery partners at full capacity to increase production. In Q4, NIO Inc. brand's overall capacity target is also set at 25,000 units monthly.

For the Firefly brand, capacity is also steadily improving. We hope its peak capacity can reach around 6,000 units monthly in Q4. Combined, the three brands' peak capacity can reach approximately 56,000 units, supporting our goal of 50,000 monthly deliveries.

Considering current capacity needs to prioritize existing strong demand, we won't have new model deliveries this year. The originally planned ONVO L80 for this year won't be scheduled for delivery due to lack of additional capacity support; whether to launch will be further observed based on market conditions. For subsequent product planning, besides ONVO L80, NIO Inc. will launch two large SUVs next year, including ES9 and the large five-seat SUV ES7. Including L80, three large SUVs will achieve delivery next year. This year's NIO Day will focus on the new ES8 launch, combining with ES8's launch rhythm as the event centerpiece.

**Tim Hsiao**: My second question concerns pricing strategy. We noticed both L90 and new ES8 adopted highly competitive pricing strategies at launch. Will this pricing strategy apply to all subsequent new models? Also, how should we view profit margins for next year's new models? When new models are gradually launched next year, what profit margin level would be more sustainable and achieve reasonable balance?

**Stanley Qu**: Our long-term core goal is achieving 20% comprehensive gross margin, with clear gross margin planning by brand: NIO Inc. brand should progress from 20% toward 25%, ONVO brand needs to maintain above 15% and strive higher, while Firefly brand targets around 10%. For the launched ES8, L90, and new models to be introduced next year, we made full preparations during product definition, striving to support competitive pricing with solid cost control. Therefore, no need to worry excessively - compared to previous generation products, this generation has sufficient cost advantages, backed by technological accumulation, self-research breakthroughs, and various cost control measures achieving cost reduction effectiveness.

**CICC Analyst Jing Chang**: My first question still focuses on L90 and ES8. Currently, we've seen these two new models demonstrate stronger product capabilities, highly competitive pricing, and maintain healthy gross margins. Management has previously introduced technology and platform upgrade situations. Beyond this, can you share more about the success factors behind these two new models? For example, adjustments or optimizations in supply chain, dealer networks, etc.?

**William Li**: Indeed, our new generation products have very strong competitiveness - comprehensive competitiveness. From product capability and cost competitiveness perspectives, the core still comes from technological innovation. For example, our 900V high-voltage platform enables highly integrated electric drive and various high-voltage components, with significant results in lightweighting and performance optimization, while lightweighting technology itself brings obvious cost advantages. This integrated design and technological innovation both enhance product experience competitiveness and strengthen cost competitiveness. Everyone can see L90 and new ES8's abundant front and rear trunk storage space, benefiting from this integrated R&D capability. In intelligence, the new digital architecture (central computing unit + domain controllers) is also highly integrated, supporting rich intelligent functions while having advantages in cost and weight control. Take smart fuses as an example - their integration has dramatically improved: previously, vehicle fuses required nearly 10kg and over 8 liters volume, now integrated into the mainboard, not only enabling fine management of vehicle power consumption but also optimizing experience and reducing costs. These examples clearly show technological innovation both enhances user experience and brings cost competitive advantages. Take chips as an example - despite large R&D investment, our flagship chip 9031's single-chip performance matches four flagship autonomous driving chips in the industry, with significantly lower costs. Though upfront R&D expenses are high, ultimately reflected in the bill of materials (BOM), overall costs drop substantially.

Regarding technology roadmap, I want to particularly emphasize NIO Inc.'s "chargeable, swappable, upgradeable" technology roadmap. We can match users with the most suitable battery packs, while industry peers commonly use lithium iron phosphate to create oversized battery packs (some even reaching 90-100kWh) to improve range. Though lithium iron phosphate controls costs, it causes dramatic battery pack weight increases. In comparison, our 102kWh and 85kWh battery packs are about 200kg lighter than industry battery packs with 600-700km range - ONVO's 85kWh battery pack weighs only 400+kg, NIO Inc.'s new ES8's 102kWh battery pack weighs only 500+kg; while industry 90-100kWh battery packs basically weigh 600-700kg. This difference reflects our "chargeable, swappable, upgradeable" technology architecture's significant advantages in weight control and cost optimization.

From automotive industry product competitiveness perspective, including experience and cost, it basically depends on three levels of capability. First is technology roadmap, which I just discussed NIO Inc.'s many advantages; second is product planning, clearly targeting which markets and providing which products for users; third is product definition. From self-summary and reflection, we're clear in direction and perform well in technology roadmap and product planning - such as serving different users through multiple brands, persisting in chargeable-swappable-upgradeable models, and ONVO's full-stack R&D. But in product definition, we indeed accumulated many lessons in the past. Fortunately, new generation products like L90 and ES8 learned from industry best practices and corrected past shortcomings. With competitive technology roadmap, reasonable product planning, plus product definition fitting Chinese user needs, these two models finally achieved current market welcome.

In supply chain, establishing long-term win-win relationships with partners is crucial for enhancing cost competitiveness. Over the past year or two, we significantly adjusted supply chain strategy, focusing more on deep binding with partners who recognize our technology roadmap and are optimistic about long-term development, promoting "partner-centric" supply chain designation strategy. Subsequently, in new platform supply chain designation, we'll more resolutely practice this concept, jointly setting cost targets and multi-dimensional collaboration targets with partners, strengthening cost advantages through bilateral cooperation.

**Bank of America Analyst Ming Hsun Lee**: I also have two questions. First, can you confirm the company's 2026 new model planning? I want to confirm whether there will be at least five new models launched, including ES6, ES7, ES9, L80, and Firefly brand's second model?

**William Li**: For 2026 product planning, we previously mentioned mainly launching three large new vehicles. As for the four models ET5, ET5T, ES6, EC6 that just completed renewal this year, there are no new model plans next year. These four models have upgraded intelligence systems to Cedar S NIO Inc. central computing platform this year, with chips and operating systems fully switched to the new platform, and we recently announced standard 100kWh batteries, further enhancing range competitiveness. We believe they maintain strong competitiveness in the foreseeable future, and these products' interiors and exteriors have also been significantly upgraded. However, small version updates like the recently launched champion commemorative edition belong to regular product adjustments and will continue next year. Additionally, Firefly brand has no plans for a second model.

**Ming Hsun Lee**: My second question concerns operating expense control. In 2026, what level will quarterly R&D expenses maintain? Can quarterly non-GAAP R&D expenses be controlled around RMB 2 billion? Also, can you introduce the latest capital expenditure plans for 2025 and 2026?

**William Li**: This year we've done extensive efficiency improvement work in R&D. Based on basic operating unit mechanisms, the company's R&D efficiency has improved significantly, and return on investment has substantially increased. Next year we'll basically maintain RMB 2-2.5 billion quarterly R&D spending to ensure we maintain long-term competitiveness. Main changes are still in vehicle models, while basic R&D investment is at reasonable levels.

**Stanley Qu**: Regarding next year's capital expenditure, current next year's operating plan hasn't been formulated, so specific scale can't be projected yet, but I can share two directions: First is battery swap station construction, where we still hope to leverage social resources to build more stations; second, overall capital expenditure will be affected by new model launch and development rhythm, but preliminary plan is maintaining levels similar to this year, and we'll strive to control lower than this year if possible, which also depends on new model development rhythm and capacity advancement.

**UBS Analyst Paul Gong**: My first question concerns the impact of all new brands fully adopting 100kWh batteries. Can you share the financial impact of this measure? Obviously, we can see 100kWh batteries improve vehicle competitiveness, but for the company, how much additional cost does this bring?

**Stanley Qu**: The impact of standard 100kWh batteries was mentioned when we announced the policy and communicated with users. When 2025 5566 models launched, we had many promotional activities. This price adjustment will recover corresponding promotions, such as RMB 38,000 discounts, converting to standard pricing. But from users' final transaction price (TP) perspective, there's basically no significant change, and no obvious impact on 5566 models' transaction prices and gross margins. I should add that after standard 100kWh batteries, this has already provided considerable help to our sales openings, though this sales policy just implemented for a few days, long-term impact still needs observation, but currently looks like positive change.

**Paul Gong**: My second question concerns switching to self-developed chips' impact. I think we didn't mention self-developed chips saving costs just now, and due to cost-volume linkage, cost savings also depend on volume. Can management provide specific information, for example, if delivering 20,000 new vehicles monthly with self-developed chips, how much cost can each vehicle save? If monthly deliveries reach 50,000, what positive impacts would there be from cost savings perspective? Since switching to self-developed chips, we hope for more accurate estimates and sensitivity analysis.

**William Li**: From chip cost perspective, everyone knows our R&D expenses are recorded as current period costs, so current chip volume doesn't strongly correlate with per-chip prices. When we purchase from chip foundries, we're charged by wafer, which doesn't have strong direct relationship with our vehicle deliveries. Comparing performance, versus our second-generation chips, costs have dropped dramatically for equivalent computing power, and even compared to current externally sourced, performance-equivalent global flagship autonomous driving chips, costs have also decreased significantly. Though I can't disclose specific cost data, the reduction is very significant.

**HSBC Analyst Yuqian Ding**: First question wants to discuss pricing with management. ES8 and L90 pricing are both attractive with good sales. How will management evaluate potential internal cannibalization of existing product portfolio, such as ET5, ET5T, ES6, EC6, or L60, and potential impact on next year's new model product line?

**William Li**: First, pricing needs comprehensive consideration of market, cost, and price-volume relationships. After L90 launch, it had positive impact on L60 sales - August L60 orders hit this year's high, growing substantially versus July. Currently users still need to wait for order-to-delivery (OTD), sufficient to prove this positive effect. New ES8 launch, combined with recent standard 100kWh batteries for 5566 models, we've sorted out clear pricing system. Future brand momentum improvement will also positively impact 5566 sales - we've already observed some positive market signals, though there might be slight fluctuations initially due to frontline sales adaptation issues, but long-term outlook is definitely favorable.

I also need to emphasize that recent hot sales of L90 and new ES8, plus multiple quality pure electric large three-row SUVs (some from extended-range enterprises) in the market, greatly influenced user mindset. From data, pure electric (BEV) grew 39% year-over-year in H1, extended-range (RE) grew 14%. If including July-August data, the gap will widen further. Obviously, medium-large and large pure electric SUVs' competitiveness is gradually emerging, and the "pure electric large three-row SUV era" is approaching. This market change's user mindset shift will also positively drive sales of existing products like ES6 and L60.

**Yuqian Ding**: Second question will further explore operating expenses. Management previously mentioned technological innovation, design optimization, and R&D investment content. Can you provide more quantified information and detailed breakdown? For example, are there clear operating expense cost targets, and can you provide more detailed breakdown of various cost optimization measures?

**Stanley Qu**: Under the background of achieving non-GAAP break-even in Q4 this year, our R&D expense planning is: without affecting mainline output and long-term competitiveness, control around RMB 2 billion; sales and administrative expenses (SG&A) hope to control within 10% of total sales revenue, which is our key target for achieving Q4 break-even. Long-term perspective, as mentioned in previous 2026 planning, we hope quarterly R&D expenses maintain in RMB 2-2.5 billion range, specifically adjusted according to new product launch rhythm; for SG&A expenses, we'll continuously advance optimization toward higher efficiency targets.

**Goldman Sachs Analyst Tina Hou**: From long-term perspective, how should we view L90 and ES8's average monthly stable sales?

**William Li**: China's current automotive market competition is indeed extremely fierce, and from intelligent electric vehicle sales trends, new car launches often have high initial popularity, then gradually enter stable sales phase. In the market, it's actually very difficult for a model to maintain high sales long-term, and currently there are few such cases. Therefore, we can hardly precisely define these two models' long-term steady-state sales specific values, but we'll definitely work toward higher targets. Starting this year, NIO Inc. and ONVO teams have been building new marketing paradigms, with core goal of extending new car popularity cycles as much as possible, helping models maintain relatively high sales levels even in stable sales phases. Objectively speaking, this process faces considerable challenges, and we currently can't immediately judge whether this marketing paradigm can help competitive products like L90 and ES8 reach long-term stable sales levels satisfactory to market, investors, and users. More time is needed for observation and verification.