On September 2, Nio reported second-quarter earnings showing a 71.2% quarter-on-quarter jump in deliveries to about 72,000 vehicles. Revenue rose 57.9% to RMB 19.01 billion (USD 2.7 billion), while net loss narrowed 26% to RMB 4.99 billion (USD 700 million).
Nio’s main brand led with 47,132 deliveries, followed by Onvo with 17,081 units. Firefly, which only began shipping on April 29, added 7,843 vehicles.
Following the August launch of the Onvo L90, the sub-brand quickly became Nio’s second growth engine. Onvo delivered 16,434 units that month, lifting total sales nearly 50% to 31,305 vehicles, including 10,525 under the Nio brand.
Starting in October, Contemporary Amperex Technology (CATL) batteries will be installed in the Onvo L90, a move expected to further support deliveries.
During the earnings call, Nio founder and CEO William Li said production capacity for the Onvo L90 would reach 15,000 units per month in October and increase to 25,000 units per month in the fourth quarter. He added that the new Nio ES8 is expected to hit a monthly output of 15,000 units in December.
Li attributed the L90’s early traction to stronger organizational execution and a shift in product definition. For instance, he said customer requests for a rear-seat entertainment screen and larger refrigerator led to the inclusion of a panoramic second-row display and oversized cooler. “If we’re doing it, we should do it thoroughly,” he said.
The L90’s popularity prompted Li to delay the launch of the Onvo L80 from this year to next, while also confirming that more large SUVs will debut in 2026. He emphasized that with limited resources, Nio is prioritizing product depth over breadth.

Price and profitability pressures
Despite higher volumes, Nio’s average selling price declined. In the second quarter, even with deliveries of the high-end ET9, the average vehicle price fell RMB 12,000 (USD 1,680) from the prior quarter to RMB 224,000 (USD 31,360). The drop underscores the challenge of balancing brand positioning with sales growth.
To spur demand, the company cut prices aggressively. In the third quarter, the Nio “5566” series (comprising the ES6, EC6, ET5, and ET5T) came standard with 100 kilowatt-hour batteries, while the new ES8 was priced more than RMB 110,000 (USD 15,400) lower than its predecessor. Li said all future models will be priced relative to the new ES8, which serves as a benchmark.
This strategy helped Nio’s deliveries exceed 30,000 units in August for the first time. Preorders for the new ES8 are reportedly stronger than those for the Onvo L90, with expectations that monthly sales will surpass 10,000 units once deliveries begin.
Li set an ambitious target: if Nio can achieve monthly deliveries of 50,000 units in the fourth quarter, raise gross margin to 17–18%, hold sales expenses at about 10% of revenue, and keep R&D spending to 6–7%, the company could post its first non-GAAP profit. That implies quarterly revenue of around RMB 30 billion (USD 4.2 billion), with sales expenses capped at RMB 3 billion (USD 420 million) and R&D expenditure between RMB 1.8–2.1 billion (USD 252–294 million).
Margins, however, remain thin. Q2 gross margin rose just 0.1 percentage point from Q1. Li said the impact of new models had not yet been reflected, since the “5566” lineup completed its platform upgrade only in late May and made up just 20% of deliveries. By Q4, he expects per-vehicle margins of 16–17%, with Nio-branded cars at 20–25%, Onvo at 15%, and Firefly at 10%. The Onvo L90 and new ES8 are targeted at 20%.
Cutting costs and tightening focus
Nio has begun to curb spending. In Q2, R&D costs fell to RMB 3.01 billion (USD 420 million), or 15.8% of revenue, down from RMB 3.18 billion in Q1 and RMB 3.22 billion a year earlier. The company has scrapped its smartphone iteration plan, slowed updates on its NT2.2 and NT2.5 platforms, and merged teams across Nio, Onvo, and Firefly to cut duplication.
Selling, general, and administrative expenses also declined to RMB 3.97 billion (USD 560 million), representing 20.9% of revenue, down from RMB 4.4 billion in Q1. Li acknowledged the ratio remained higher than desired but said efficiency gains would bring it down further. He cited Firefly’s rollout as an example, noting that just 31 new sales staff supported average monthly deliveries of 4,000 cars.
Lessons from past crises
In a media interview, Li contrasted today’s challenges with the company’s 2019 downturn, when Nio was rescued by local government funding in Hefei. “In a way, we got through it by luck,” he said. “It’s different this time round. We fell into a deeper, longer pit. The process of climbing out has forced us to better recognize our competitive advantages and organizational problems.”
He stressed that profitability would validate Nio’s trajectory. “Financial losses affect many things, be it our user conversion rate, our hiring, or our supplier relationships,” he said. “It’s very important to prove that we can turn a profit. It’s not about proving it to others. It’s that we ourselves are very clear that, at this point in time, we need to do this to validate our progress.”
Onvo’s turnaround has been led by Shen Fei, formerly head of Nio’s energy business. Li complimented Shen’s ability to establish systematic processes and maintain strategic focus while adapting execution.
Globally, Li admitted Nio overspent on its expansion into Europe in 2021, when it exported its full ecosystem, including battery swap stations and the Nio House project. “The costs were too high, and we underestimated geopolitical risks and energy price fluctuations,” he said. Now, the company is leaning on local partners and lower-priced brands like Firefly and Onvo for overseas growth.
On the battery swap network, Li said profitability will take time. A station requires about 60 daily swaps to break even, but the current average is around 30. He added that the network nonetheless strengthens sales and brand value.
Li framed Nio’s path to profitability as a test of maturity:
“In 2019, luck helped us survive. This time, it’s about proving we have the skills to do it ourselves.”
The following transcript has been edited and consolidated for brevity and clarity.
Q: How is this crisis different from the one in 2019? What feels different for Nio going through these two downturns?
William Li (WL): This time, climbing out of the bottom depended on our improved capabilities. In 2019–2020, we survived partly because of user support and partly because the Hefei and Anhui governments extended a helping hand. In a way, we got through it by luck, and we ourselves didn’t really improve much in that process.
It’s different this time round. We fell into a deeper, longer pit. The process of climbing out has forced us to better recognize our competitive advantages and organizational problems. You could say we’ve actually started to build some real skills.
Q: What does it mean for a company to turn a profit in a given quarter? If Nio becomes profitable in Q4 this year, does that determine whether the company has a future?
WL: Financial losses affect many things, be it our user conversion rate, our hiring, or our supplier relationships. The impact is everywhere.
So at the right moment, it’s very important to prove that we can turn a profit. It’s not about proving it to others. It’s that we ourselves are very clear that, at this point in time, we need to do this to validate our progress.
Q: Some of your peers are spending heavily on GPUs to train large AI models. Will that be a major expense for Nio in the future? Will the Nio smartphone still iterate annually?
WL: The smartphone business is something we’ll come back to when we have more money.
On the question of GPUs, in the past we bought a lot of cards too. But later we realized that more cards don’t necessarily mean better models, and DeepSeek proved that.
One advantage of having less money is that it forces you to focus on what’s essential. When you have lots of money, everyone thinks about spending. With less, you have to prioritize projects and ask whether they’re worth doing, and what the output will be. The same applies to GPUs.
Right now, we’re managing GPU budgets very tightly. We rent many cards. When we need them, we rent, and when we don’t, we return them. Generations of cards were being run without people even knowing what was being calculated. That’s why we need to manage it closely. Not having as many GPUs doesn’t mean you can’t produce good results. Everyone has to run the numbers.
Q: Behind the success of the Onvo L90 and the new ES8, what adjustments has Nio made in product definition?
WL: In the past, Nio tended to bury value inside the product, so the explicit value wasn’t strong enough. With the third-generation ES8, we corrected that.
We also used to define products based too much on our own preferences. For example, with the entertainment screen facing the rear seats, I was always in conflict with my kids at home about screen time, so I didn’t want them watching in the car. But as they grew up, I realized I couldn’t control it. Kids want to watch in the back seat. So I compromised: if we’re going to have a screen, let’s make it really good.
The refrigerator is similar. We used to doubt whether demand was that strong. But we found that people did want it, so we went for a large cooler. If we’re doing it, we should do it thoroughly.
Q: What lessons has Nio learned from the success of the Onvo L90 and the new ES8?
WL: Sometimes a product may not sell well at first, but as organizational capabilities improve, sales will rebound to where they should be. The Onvo L60 is a good example.
It’s actually an interesting car. This year, it’s been consistently in the top three among pure electric SUVs priced RMB 200,000–300,000 (USD 28,000–42,000). It drew criticism at the start of the year, but sales have been improving, which shows the product itself is competitive. With stronger organizational capabilities, sales can return to their rightful level.
I believe product strength determines the sales ceiling, but organizational capacity in areas like pricing, sales, channel management, and marketing will determine how many units a car can actually sell. Sometimes if sales don’t meet expectations, it’s not the product’s fault.
I often tell our team that the Onvo L60’s product definition is solid, but other capabilities weren’t keeping up. That’s why this year we restructured our marketing system extensively. Between April and May, headcount was cut by 40%, but sales rose 40%. That reflects deep changes.
Right now, Nio’s mindset is to play the long game. Over the past two years, from myself to frontline staff, we’ve all built a mindset of focusing on the work and not getting distracted.
Q: How has Nio prepared production capacity for the Onvo L90?
WL: The biggest change is adapting to the delivery cycle of smart electric vehicles. A new model now has a bigger launch effect than before: demand surges at release, then dips, and later stabilizes. The supply chain has to match this pattern.
It’s similar to consumer electronics. Smartphones need inventory prepared in advance to enable rapid deliveries. So this time, we started stocking earlier and in larger volumes.
Of course, this comes with the need to tie up capital. Preparing 3,000–5,000 units requires RMB 1–2 billion (USD 140–280 million). Some supply chain support can be arranged through partners, but some components need to be stocked ahead of time. That’s where we’ve changed.
Q: Shen Fei was previously in charge of Nio’s energy business. Why did you choose him to lead Onvo?
WL: To build a successful brand and product, you need systematic capabilities. When Shen took over, Onvo was in a tough spot from facing lots of negative press and low morale. We knew it needed a proper system.
Shen is excellent at building systematic capabilities. Our battery swap system, with its chargeable, swappable, and upgradeable features, is something he did very well. That’s why we asked him to lead Onvo.
He also has a strength in sticking to the right direction without wavering, while still being very thoughtful in execution. That’s exactly the kind of colleague I need.
Q: What are the upgrade plans for models like the “5566” series, EC7, and ET7 next year? Does Onvo plan to launch A- or B-segment sedans?
WL: Next year, we’ll complete the lineup of larger vehicles. The “5566,” EC7, and ET7 won’t get upgrades. We just finished upgrading the “5566” with a new intelligent system and made 100 kWh batteries standard, which improved range. The “5566” also uses our third-generation design language, with horizontal screens inside. That’s a big enough change.
Beyond next year is too far out, though of course we have internal plans. For Onvo, the focus is on completing the SUV lineup, which we believe can already support solid sales. Other plans won’t come until after 2026.
Q: The new Nio ES8 seems to be cannibalizing demand for the ES7. Will the ES7 take the route of adding features and cutting prices?
WL: The new ES8 is a benchmark model. The first-generation ES8 was also a benchmark as it had the highest sales and best margins. The second-generation ES8 deviated, with costs going over RMB 500,000 (USD 70,000). Now we’ve returned to having a benchmark model. All future product pricing and technical platforms will be based on it. That’s also why we’ve made 100 kWh batteries standard on the “5566” series.
We also have strong cases in intelligent engineering. For example, with the new ES8 we’ll launch a “4D” navigation feature. When the system detects an incline, the sensors feed information to the CDC (continuous damping control) suspension, precharging it. Without a vehicle-wide operating system, that wouldn’t be possible. The effect is unique in the industry.
Q: What are Nio’s thoughts on globalization?
WL: Learning about globalization has been costly. We entered Norway and Europe in 2021 because high-end cars sell mainly in China, the US, and Europe.
At that time, we exported our entire system, including more than 60 battery swap stations, Nio House locations, and direct sales. But we underestimated factors such as geopolitics and rising European energy prices.
In 2021, we performed well in Norway, which encouraged us to push harder in 2022. Looking back, the costs were enormous. We later realized globalization cannot simply copy China’s model. It’s too expensive. This year, we have shifted to rely more on local partners.
Unlike before, we now have more affordable models under Firefly and Onvo to support global expansion.
There have also been surprises. For example, the ET5 Touring and Firefly were originally designed primarily for Europe, but the ET5T turned out to be a standout hit globally. Firefly also exceeded expectations in China, where we thought 3,000 monthly sales would be strong, but it has been selling more.
Q: Will battery swap services reach large-scale profitability next year?
WL: Large-scale profitability will take longer. A swap station can make money if it serves about 60 swaps per day. Right now, about 20% of our stations reach that level. On average, we’re at just over 30 per day, so there’s still a gap.
But swap stations also help us sell cars. This year, we introduced a system where regional companies can co-invest in new stations. The return on investment can be calculated: if a station helps sell 10–20 cars per year, it breaks even.
There’s also brand value, like our 318 project, which is funded by the brand team.
Overall, before building a new swap station, we now forecast the ROI from high to low, and build accordingly. Many stations already have ROI greater than three to four.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This article was written by Fan Shuqi for 36Kr.