- NIO shifted emphasis from pure volume targets to prioritizing operating profit and revenue quality, reflecting pressure to lift margins and capture higher-end buyers from legacy luxury marques.
- NIO delivered 107,658 vehicles in Q2 2026, revenue RMB 32.137 billion (about $4.77 billion), vehicle gross margin 18.5%, average transaction price RMB 406,000 (about $60,300).
- Targets the high-end luxury segment against Mercedes-Benz, BMW and Audi, but execution risks include sustaining margins, meeting ES8 delivery cadence, market sentiment and capital-market valuation volatility.
Three days after releasing its second-quarter earnings report, NIO held a small-scale media briefing at its headquarters in Caohejing, Shanghai, yesterday afternoon. The session lasted a full three hours.
NIO CEO William Li and President Qin Lihong addressed questions from the media one by one, covering everything from financial data and industry trends to the company’s second growth curve, the Germany luxury trio of Mercedes-Benz, BMW and Audi, AI, and the development of ONVO. The amount of information shared was substantial.
Li appeared relaxed. Before the briefing began, he joked, “The materials they prepared for me, we’ll digest them together. I haven’t read them yet. It’s fine, more or less.”
Neither Li nor Qin shied away from discussing valuation or the share price.
Li even joked that if he could explain why NIO’s stock had risen or fallen, he would have made a fortune long ago.
After listening back to shorthand notes late into the night, ChinaEV Home decided to capture the core of the briefing through seven keywords: surpassing Germany luxury trio, undervalued spots, the three-way AI shock, Li’s obssession to numbers, ONVO moving upmarket, the road ahead, and the EV endgame.
So, based on those seven themes, how do Li and Qin describe the future of an automaker?
Beyond Germany Luxury Trio
Li did not spend much time repeating the earnings report.
The key figures are worth briefly restating: NIO delivered 107,658 vehicles in the second quarter of 2026, up 49.4% year on year, while revenue rose 69.1% to RMB 32.137 billion ($4.77 billion).
Vehicle gross margin was 18.5%. On a GAAP basis, the company still posted an operating loss of RMB 347 million ($51.5 million) and a net loss of RMB 528 million ($78.4 million).
Excluding share-based compensation and other items, NIO reported a non-GAAP operating profit of RMB 207 million ($30.7 million) and an adjusted net profit of RMB 26.1 million ($3.9 million).
When discussing those earnings figures, Li placed particular emphasis on operating profit, which he said is the most important indicator of a company’s operating quality.
“Ultimately, when you look at a company, it’s the operating profit. If you take all the money on your books and put it in the bank — or, bluntly, put all your cash into wealth-management products — then our profit would already be very large. But what would be the point? Or if we just invested everything in wealth-management products, what would be the point?”
From an operating-profit perspective, Li said, NIO has demonstrated an ability to consistently turn technological innovation, product development and its service system into profits.
That, he said, is the foundation for healthy and sustainable growth.
Put another way, he said:
“In the past, the sequence was volume growth, revenue growth and then profit growth. Today, we need to look at the order differently: profit growth, revenue growth and then volume growth.”
Indeed, over the past year or so, the company has stopped setting specific sales-volume targets internally.
That is another way of explaining the idea of working backward from the end goal.
In numerical terms, NIO’s revenue reached RMB 57.670 billion ($8.57 billion) in the first half of this year, up 85.8% from a year earlier, while gross profit surged 282.2% to RMB 10.766 billion ($1.60 billion). Deliveries increased 67.4% over the same period.
Put bluntly, revenue is growing faster than vehicle sales, while gross profit is growing faster than revenue.
As he elaborated, Li smiled and half-jokingly, half-seriously asked the more than 10 media representatives in the room whether the media could create a new ranking.
NIO could not do it itself, he said, because the industry has traditionally ranked automakers by sales volume while rarely ranking them by revenue, total gross profit or operating profit.
Volume matters, but volume growth alone does not constitute healthy growth unless it is accompanied by corresponding increases in revenue and gross profit.
Interestingly, during the subsequent Q&A, one media outlet actually ran the numbers and said that based on revenue, NIO had already become one of the top three Chinese automakers in the domestic market, behind only BYD and Geely.
Li added that if only pure-EV sales were counted, he estimated NIO and Tesla would rank first and second.
Another key measure of operating quality is average transaction price.
Li said the NIO brand’s average transaction price was RMB 406,000 ($60,300) in the second quarter and rose further to RMB 434,600 ($64,600) in July.
That already puts NIO above the average transaction prices of Mercedes-Benz, BMW and Audi.
It suddenly brought to mind a video Qin Lihong had apparently posted earlier, showing that NIO had overtaken Mercedes-Benz, BMW and Audi in sales simultaneously across 42 cities.
If you remember, Li said a long time ago that he hoped NIO could eventually secure one-third of China’s high-end luxury market alongside the Germany luxury trio.
That day appears to have arrived earlier than expected.
The latest-generation ES8 can be seen as a symbol of that shift. As of Aug. 21, the pure-electric SUV, which starts at more than RMB 400,000 ($59,500), had completed 140,000 deliveries in 335 days.
From January through July, the ES8 ranked first in cumulative sales in both the large-SUV segment and the market for vehicles priced above RMB 400,000 ($59,500).
Li said cumulative ES8 deliveries are expected to reach 150,000 in September.
What the Market May Still Be Missing
After NIO released its earnings report, its stock price fell sharply.
So the question was inevitable. Yesterday afternoon, one reporter directly asked Li how he viewed NIO’s current share price and whether he believed the company was undervalued.
“All pricing in the capital markets is a market price. It reflects what everyone thinks about you at a particular point in time. It incorporates many things. Just because our market capitalization was very high in 2021 doesn’t mean the company was several times better than it is today. NIO is definitely much healthier today than it was in 2021. If I could clearly explain why the market is pricing it this way, wouldn’t I have made a fortune already?”
Li gave that answer, but when asked why the stock had behaved this way, he frankly admitted that he could not explain it either. If he could, he joked, he would have made a fortune long ago.
Li remained relaxed and did not shy away from the subject, even sharing a small anecdote.
That very morning, he had been in the same room training this year’s new hires, the “Spark” graduates.
Because NIO is employee-owned, he joked with the young employees that it was actually a good time to join. With the share price low, they would receive more shares as part of their compensation.
It is hard to know quite how to evaluate that joke.
But when the discussion turned to what aspects of NIO might be undervalued, Li became serious.
The first was NIO’s 12 fully self-developed technology stacks.
He said the long-term competitiveness of NIO’s full-stack capabilities — spanning chips, operating systems, AI and battery swapping — is clearly underestimated.
From the perspective of full-stack capabilities, there is substantial room for expansion. It is essentially infrastructure for the AI era. From a business-expansion perspective, it offers considerable potential.
“But there is no need to worry about this.”
The second potentially undervalued asset, Li said, is NIO’s brand scarcity in the era of intelligent electric vehicles.
At the briefing, he asked the media: Outside Tesla and Toyota, which automaker has the most valuable brand in the global market?
After waiting for a moment, he gave the answer: Ferrari, with a market capitalization of $70 billion, approaching $80 billion.
Li argued that this shows how important brands remain regardless of how the industry changes, because brands are tied to emotional experiences and to consumers’ deeper emotional perceptions and associations.
In his view, among vehicles priced above RMB 300,000–400,000 ($44,600–$59,500), there are very few brands in China that can truly compete with NIO on brand appeal.
The implication was that there may be few, if any, such brands in the Chinese market.
Li said NIO’s systematic competitiveness in branding should be among the strongest in the market.
A brand represents long-term, stable competitiveness and sustainable profitability. It cannot be built through a breakthrough in just one area.
As for how much time investors would need to evaluate the company, Li said he could not make decisions on behalf of investors.
NIO’s job is to execute well, continue investing in technological innovation and brand building, and maintain strategic discipline.
Whether it is the share price or valuation, he said, both will ultimately return to the company’s operating performance and profitability.
The final potentially undervalued area is the service and community revenue generated as NIO’s installed user base grows.
Li disclosed that, thanks to the distinctive economics of its growing installed base, NIO’s service and community revenue exceeded RMB 10 billion ($1.49 billion) last year and reached RMB 5.8 billion ($860 million) in the first half of this year.
“This is a business model that other companies simply do not have at the moment.”
The Three-Way AI Shock
At this point, it is important to record Li’s description of the three major shocks brought by the AI boom. The automotive industry’s current undervaluation is directly related to them.
The first is cost.
AI data centers are rapidly increasing demand for memory and PCBs while also competing for resources such as copper, aluminum and batteries.
“You might ask, what did we have to do with AI data centers before? Well, sorry, memory capacity is now becoming tight, and even PCBs could be in short supply. Why? Because those PCBs are going to AI. Some customers over there are willing to pay five or even eight times the price to buy them.”
When data centers are willing to pay several times more, Li said, suppliers naturally shift capacity toward them.
Higher prices are therefore inevitable, and the impact is “very real.”
According to figures Li gave at the briefing, NIO’s material cost per vehicle in the second quarter was about RMB 14,000 ($2,080) higher than at the end of 2025. Costs are expected to rise by another RMB 3,000 ($450) in the second half, bringing the cumulative increase to RMB 16,000–17,000 ($2,380–$2,530).
Memory accounts for about RMB 10,000 ($1,490) of the increase, batteries for roughly RMB 3,000–4,000 ($450–$590), while materials such as copper and aluminum are also becoming more expensive.
The second is talent. Autonomous-driving and algorithm specialists are being drawn away by startups focused on embodied intelligence.
“Some of these companies have only a handful of employees but are already valued at RMB 5 billion ($740 million) or RMB 10 billion ($1.49 billion). What are you supposed to do? Even when we were at our most aggressive, we never reached that level of madness.”
Even though these companies may not know when they will generate meaningful autonomous-driving revenue — and their revenue is “way, way earlier” than that of an automaker selling cars — “if I were an algorithm engineer, I’d be very tempted too. At least the paper wealth is real.”
The third is attention.
Capital-market attention has shifted toward AIGC and embodied robots. An automaker that does not talk about robots can seem to have little imagination left.
“Actually, the problem is not that our products and services lack imagination. It’s that many people lack imagination when it comes to our business. It’s not that our business has no room for imagination.”
Li was clearly venting some frustration.
To illustrate his point, he cited Nvidia CEO Jensen Huang’s description of the five-layer AI stack, with energy as the first layer.
“We use tens of billions of kilowatt-hours of electricity a year. Isn’t that edge computing and distributed energy?”
Li also disagrees with the current tendency for automakers to move aggressively into embodied intelligence.
Later in the briefing, he returned to the subject and said NIO has no short-term plan to enter the embodied-intelligence business directly.
Although the company is making investments in the field, its approach differs from others: it has chosen to participate through strategic investments.
When asked whether he worried that the embodied-intelligence companies NIO invests in could eventually become competitors, Li gave an unexpected answer:
“Wouldn’t that be great? If it really becomes a competitor, that means the business has succeeded.”
Li then asked rhetorically whether that was not the logic of it, before adding, “Of course, we don’t want to scare Shaoqing. We currently have no short-term plan to enter the field ourselves.”
As for how to convince investors, Li said:
“Investors? How am I supposed to convince investors? It’s difficult for me to convince them too. To be honest, the only way is to show better operating results and let time do the talking.”
Li’s Obessession to Numbers
At previous NIO media briefings, Li has typically spent much of his time discussing products, technology and users.
This time, however, he displayed an unusually strong interest in the numbers.
“I’m particularly happy running the business now. I actually enjoy it. Why? Because it’s a process of getting to the truth. Take ROI, for example. It’s not easy even to work out the full cost, let alone explain the full return. But the process of figuring it out is just like writing a good article or making a good video. It’s interesting and very satisfying.”
Even when discussing NIO’s native-AI organization, he approached the subject from this perspective.
He said NIO now retains only one core operating metric internally: EBIT.
Around that metric, the company has established a CBU mechanism that continually breaks the business down into smaller operating units: individual vehicle models, individual stores, individual Fellows, individual marketing events and even individual components.
“We’re now tackling CBU at the component level.”
A vehicle can be broken down into 17,000–18,000 individual components. Once fixed costs, variable costs and returns can be calculated for every component, every store and every event, Li said, the company can determine the optimal solution for each activity.
It is very similar to prompting. AI technology is already here; the hard part has always been knowing what you want and articulating your requirements clearly.
Qin cited Yang Bo, NIO’s head of nationwide user operations, as an example. A traditional company might view him as a national sales chief managing different regions.
Under NIO’s new definition, his first responsibility is to support 6,000 Fellows. He is “the Fellow of all Fellows nationwide.”
AI here does not mean simply adding another department. Instead, it breaks knowledge, costs and operating objects down into the smallest units that can be calculated, then rebuilds processes, performance assessments and resource allocation around them.
Another topic related to the numbers is that cost discipline can easily be interpreted as retrenchment.
“But with RMB 2 billion ($300 million) today, we can do the work that others spend more than RMB 3 billion ($450 million) on. That’s the accounting. Spend money where it matters. Spend when you should.”
Li gave two examples.
Firefly’s dedicated R&D team currently consists of only dozens of people, with much of its capabilities reused from group-level platforms. Firefly has already ranked first in the premium small-car segment for 15 consecutive months and delivered 5,852 vehicles in August.
Li said Firefly will increase production capacity in September and suggested that its monthly sales performance would be worth watching. The implication was that its order book remains strong.
The other example was the world model. Li said NIO’s cloud-computing expenditure last year was not particularly high: when training was required, the company rented computing capacity and returned it after training was completed.
Spending has increased this year, with a major version scheduled for release in the fourth quarter.
In one discussion with another automaker, Li said, the other company was shocked to learn that NIO was achieving its results without spending the tens of billions of yuan that the other company had invested.
Could CBU kill innovation? Li did not dodge that question either.
He said long-term projects can still be approved, and automotive R&D cannot reasonably be expected to pay back within six months. Time is a critical factor in ROI. Technology road maps need to be evaluated over 10 years, product planning over five years and product definition over two years. What CBU is really designed to eliminate, he said, is “pseudo-innovation” in which user value cannot be clearly articulated and priorities cannot be established.
ONVO Moves Upmarket
ONVO naturally came up repeatedly during the three-hour briefing.
For NIO Inc., ONVO is responsible for reaching a much larger market and using scale to spread the costs of NIO’s R&D, supply chain and service infrastructure.
But Li’s remarks today were particularly emphatic: the company has adjusted its expectations, and ONVO is still considered to be in an investment phase.
As for ONVO’s positioning, Li described it as occupying the upper end of Toyota and Volkswagen’s product range.
He recalled buying a CR-V many years ago, when the vehicle cost more than RMB 200,000 ($29,700) and buyers might even need to find someone to help them secure one.
The Highlander was also once a representative upgrade choice for many Chinese families.
Today, consumers who once bought higher-end Toyota or Volkswagen models may feel that returning to those brands is “not quite enough,” while buying a NIO may seem unnecessary.
Moving further downmarket, meanwhile, can feel like a compromise.
That is the market space ONVO is targeting.
“The ONVO that everyone expects to be extremely cheap simply does not exist. There is no ultra-cheap ONVO,” Li said in response to ChinaEV Home’s follow-up question.
He further explained that ONVO will not enter the price bands occupied by Toyota and Volkswagen’s lower-priced models, nor will it move into the core volume segments of Leapmotor and BYD.
The new ONVO L60 currently starts at RMB 192,800 ($28,600) for an outright vehicle purchase, or RMB 135,800 ($20,200) under the BaaS model.
Li said future products could bring the outright starting price down somewhat from RMB 190,000 ($28,200), but “not suddenly down to around RMB 150,000 ($22,300).”
Under the current long-term plan, ONVO will not move below RMB 150,000.
One additional piece of information emerged yesterday, when reports surfaced about a new ONVO product. Neither Li nor Qin elaborated on it.
As for average transaction prices, ONVO’s average transaction price from January through August this year had already reached RMB 248,000 ($36,900).
Li said there would still be a gap of “more than RMB 100,000” ($14,900) between ONVO and NIO’s average transaction prices.
Even after NIO launches its new 5- and 6-series models, the NIO brand plans to keep its average transaction price around RMB 400,000 ($59,500). For ONVO, Li said, that leaves substantial room to operate.
Li believes ONVO’s current challenges remain brand awareness and channel coverage, both of which take time to build.
Its higher-end positioning protects transaction prices and residual values, but also means it will not use sudden price cuts to buy short-term volume.
For a brand that remains in an investment phase, that requires patience and will continue to consume sales and channel resources.
Because a video of Qin at the opening of the Macau NIO House two days earlier showed him discussing the Hong Kong market and ONVO, the issue was also raised at the briefing.
Qin said he had originally not intended to disclose anything, but at the opening event, a Hong Kong customer had purchased an ES9 in Macau, while potential Hong Kong partners were also pressing him to launch operations as soon as possible.
“Relying only on Firefly launching one model would make the overall costs in Hong Kong quite high. With just one Firefly model, my main concern is that it wouldn’t make money. That could actually dampen their enthusiasm for what comes next. So we asked them to wait a little longer.”
Qin said,
“We think Hong Kong is particularly well suited to ONVO’s products. If it can enter the market together with battery-swap stations and the BaaS business model, I still believe good things are worth waiting for.”
NIO’s Road Ahead
When the briefing was held, Tesla’s Cybercab launch event was becoming the hottest story in the global auto industry that day.
ChinaEV Home therefore asked: “At a time when Robotaxis and embodied robots are rapidly drawing away capital and talent, where will NIO’s second and third growth curves come from?”
Li said, “I’m not particularly bullish on autonomous-driving taxis. At least in China, I’m not very optimistic about them. This isn’t the first time I’ve expressed this view.”
He explained that Robotaxis ultimately become a licensing business and will also face limits on how many vehicles a city can accommodate. In Shanghai, for example, deploying 10,000 or 20,000 autonomous taxis could actually make congestion worse.
As for embodied intelligence, Li does not deny its long-term potential but divides robots into three categories: performance-oriented robots that are common today, productivity robots that are already creating economic value, and general-purpose service robots that remain some distance away.
In Li’s view, intelligent driving in cars, food-delivery robots in hotels and robot vacuum cleaners at home all fall into the productivity-robot category.
They do not need to look like humans. What matters is whether they can actually perform useful tasks.
General-purpose service robots could certainly become a huge business, he said, but every company will have its own judgment about when to enter and with what products.
Automobiles were already a huge market before NIO existed; where the market for general-purpose robots will be remains an open question.
Li said:
“What I mean is that robots have a huge future. For now, we’ll just focus on selling cars. That’s basically it.”
At the briefing, he said at least twice that NIO should be allowed to “dare to be the last mover.”
Perhaps because the question kept coming up, he also said that robots could ultimately cost just RMB 30,000 ($4,500) each.
“Sixty kilograms, materials plus two smartphones — I’ll put it right here. Isn’t that what the end state looks like? What are we so anxious about?”
In Li’s view, there is still enormous room for innovation in automobiles. Even the so-called dexterous hands on robots ultimately rely on electric motors.
A car already contains around 200 small motors, so how many degrees of freedom could be created by combining them?
That is why the second growth curve Li has designed for NIO points toward recurring revenue generated from its installed user base through services and the NIO community.
In NIO’s 2025 financial report, “other sales” reached RMB 10.604 billion ($1.58 billion), while the figure was RMB 5.828 billion ($870 million) in the first half of this year.
This financial category also includes used vehicles, parts, after-sales services and technology R&D services, so it cannot be equated entirely with “service and community revenue.”
But the recurring revenue generated by a user base in the millions has clearly reached meaningful scale.
As for a third growth curve, it could come from energy or Physical AI, but that horizon is likely further out.
The area that deserves the most attention remains the automobile business.
Li explained that NIO has 12 full-stack technologies, each of which could potentially be extended into other industries. What the company can do is one question; what it should do at this stage is another.
Not long ago, Li was invited as one of the keynote speakers at a Shanghai integrated-circuit industry summit.
There were four keynote speakers in total, two of whom came from the automotive industry. Li joked on stage: Why invite an automaker to a semiconductor industry conference?
Because automakers are operating at enormous scale.
Today’s intelligent electric vehicles contain thousands of chips, large amounts of memory and software that is continuously updated. They are high-value edge-computing devices in their own right and a major real-world platform for Physical AI.
Li even said that NIO’s market share in Shanghai was 8% in the first half. If it could reach that level nationwide, NIO’s sales could triple.
If it could reach its market share across the Yangtze River Delta, sales in China alone could increase 2.5 times.
These opportunities have not yet been fully captured, and shifting the company’s focus to other industries now would amount, in his words, to “leaving the food in front of us uneaten.”
That is why Li said now is not the time for NIO to pursue embodied intelligence and similar ventures. Resources and management attention are limited, and some initiatives need to allow NIO to “dare to be the last mover.”
EV’s Endgame
For NIO to sustain annual growth of more than 40%, one of Li’s core answers is that the company is positioned in the direction of continued growth in pure-EV penetration.
In July, NEVs accounted for 60.4% of all new-vehicle sales in China for the first time, while monthly pure-EV sales exceeded 1 million units. Pure EVs, once seen as a choice for a minority of consumers, are entering the true mass market.
For NIO, the significance goes beyond simply having more potential customers.
“Sticking with pure EVs makes our operating efficiency much higher, whether in R&D or sales. Think about it: if I develop a vehicle and have to consider multiple powertrain types, I have to spend more from the R&D stage. The supply chain becomes more complicated. In sales and showrooms, you have to display another set of vehicles with different powertrains. Services, after-sales and parts all become more complicated.”
For companies that sell combustion-engine, plug-in hybrid and range-extended vehicles at the same time, growth in pure EVs must first offset the decline in legacy powertrain products.
NIO does not have that hole to fill. All of its growth is concentrated directly on pure EVs.
As the briefing drew to a close, one reporter asked Li: If you had to define China’s auto industry in 2026 with a single keyword, what would it be?
He paused and said there were still several months to go and anything could happen — perhaps the market could suddenly rebound toward year-end.
Qin, sitting beside him, added: One thing is certain: pure EVs are becoming mainstream. “We’ve recently spoken with many friends, including users, and there’s still an ingrained way of thinking that because I chose a pure EV, I must be one of the minority of early adopters. But actually, today, pure EVs are the majority.”
Li immediately picked up the point: “Yes, we can call it mainstream pure EVs — pure EVs becoming mainstream. That’s good. As expected, Lihong thinks deeply.”
Three hours of discussion — from surpassing Germany luxury trio and being undervalued by the capital market, to AI-native organizations, Robotaxis, embodied robots and the next growth curve — ultimately came back to pure EVs.
For Li and Qin, that is the answer to where the future of the automobile lies.
Discover more from ChinaEVHome
Subscribe to get the latest posts sent to your email.
