Li Bin began to calculate another account.

Author | Zhou Zhiyu

"Last year they asked us if we could make a profit. Now that we are profitable, they are asking us how we view the future."

On September 3, NIO founder, chairman, and CEO William Li discussed another type of pressure the company faces. Just the evening of September 1, NIO released a profitable financial report, marking its third consecutive quarter of positive operating profit.

In the past few years, the market was willing to give new players time; as long as sales increased, losses could be explained as efforts to seize market share, build distribution channels, and invest in technology. Now, the market wants not only profitability, but also continued growth with profits.

The AI boom is squeezing the automotive industry from three directions. AIDC is drawing away a large amount of resources such as memory, PCBs, copper, and aluminum, raising the hardware costs of smart cars; autonomous driving and algorithm talent is being diverted to AI startups; and the capital market's attention is also shifting to data centers and robotics, making the car-selling business less attractive.

“Many companies have only a few people, yet their valuations are 5 billion or 10 billion yuan. What can you do then? Even at our most frenzied, we never reached that level of madness,” Li Bin said.

This contrast is reshaping the situation for car companies. Smart cars, already large-scale commercialized physical AI, are now competing with AI companies that haven't yet generated revenue for parts, talent, and valuations. Automakers are simultaneously bearing higher costs and needing to prove themselves through operating profits.

Li Bin doesn't intend to divert resources to hotter concepts. "Robotics probably have a great future, but right now we should focus on selling cars." In his view, the automotive market itself is far from finished. To push monthly deliveries to 40,000 vehicles in the fourth quarter and strive for 40% to 50% annual growth in the third growth cycle, NIO needs to first deepen its current business.

"The meal of the car industry isn't finished yet."

NIO has basically completed its new car launch target for this year by mid-July.

Qin Lihong is "cautiously optimistic" about the auto market in the fourth quarter. From January to July this year, the market performance was below initial expectations, and a slight rebound is possible, "otherwise the whole year will be too dismal."

Even with the market recovering, NIO has no plans to suddenly boost sales.

“Sometimes when we talk about sales, we use grand narratives, and everyone feels like they’re standing on a mountaintop, and they can sell a car with a wave of their hand. That’s not how it works.”

Qin Lihong did a multiplication problem on the spot. Suppose NIO and Ledao have 10,000 front-line sales staff nationwide, each working more than 200 hours per month. Multiplying the two figures together gives the total amount of time the sales system can use in a given month.

He divided this time into "grids". Every ineffective meeting held or every useless form filled out on the front line meant less time to interact with customers.

"During the time we've been talking to you all this afternoon, many squares have been drawn out in my mind, and once they're drawn out, they can never be brought back."

NIO also used to manage sales using top-down process metrics. Lower-level staff spent a lot of time reporting, sometimes even "seriously faking it." Now, NIO hopes to return these metrics to the front lines, allowing each salesperson to generate more sales per unit of time.

More than a year ago, NIO canceled the company-wide monthly delivery assessment. Sales figures are still being monitored, and management will continue to track them: how many deliveries came from existing orders, and how many were new orders in the current month; how many staff members a store used, and how many orders it secured.

Sometimes people tend to confuse delivery volume with new orders. Li Bin said that the company sees new orders increasing month by month.

In the third quarter, NIO expects to deliver 108,000 to 111,000 vehicles, averaging over 30,000 vehicles per month. If it can stabilize at 40,000 vehicles per month in the fourth quarter, it will have already achieved a significant increase compared to the third quarter.

The 30,000 to 40,000 vehicle mark is a common platform for leading emerging automakers. In the early days, the new energy vehicle market expanded annually; a product, as long as its direction was correct, could attract a group of users away from gasoline-powered vehicles. Now, all mainstream price ranges are crowded with new energy vehicles, and automakers, to increase sales, often have to take orders from other new energy brands.

When a car is first launched, its new features and prices are quickly followed by competitors. It's not uncommon for a single blockbuster model to boost monthly sales from 20,000 to 30,000 units. To maintain sales above 40,000 units consistently, multiple models need to sell well simultaneously. New energy vehicle startups have fewer models and rely heavily on direct sales systems. When a flagship model enters the final stages of its model cycle, fluctuations quickly impact the company's overall sales.

NIO believes the pure electric vehicle market will continue to expand. In July of this year, pure electric vehicles accounted for 44.3% of all powertrain options; the penetration rate of pure electric vehicles in large SUVs has also exceeded 40%. With the gradual maturation of charging networks and battery technology, pure electric vehicles are moving from being a minority choice to entering the mainstream market.

Li Bin did some calculations using the Shanghai market as an example. NIO's market share in Shanghai has already reached about 8%. If it could reach this level nationwide, sales would have three times the potential. Even if it only reaches the level of the Yangtze River Delta region, the Chinese market could still grow by more than two times.

The automotive industry isn't finished yet, and NIO doesn't need to chase after Robotaxi and humanoid robots for the time being. Li Bin doesn't believe Robotaxi will become a large enough business in China in the short term; licensing, city capacity, and road congestion will limit its scale. As for embodied intelligence, "NIO should be allowed to dare to be a pioneer."

The pure electric vehicle market continues to expand, leaving room for NIO to grow. The remaining issues are less exciting: new models need to be launched promptly, frontline sales need to improve efficiency, and monthly sales of 40,000 units need to be transformed from a short-term target into a stable level.

Calculating the costs for each vehicle

Sales need to increase, but the cost per vehicle is rising.

“We used to say that when volume increases, costs decrease. Now, when volume increases, costs may also increase,” Li Bin said.

The procurement of memory is a prime example. In the past, the more automakers bought, the larger the discounts suppliers offered; now, some suppliers are willing to lower prices for small-batch purchases, but once customers increase their quantities, the additional parts become more expensive. Li Bin calls this situation "diseconomies of scale."

AIDC is competing with automakers for memory, PCBs, and materials such as copper and aluminum. Some customers are willing to pay several times the price to secure production capacity, so suppliers naturally prioritize selling to those who offer higher prices. "What did we have to do with the AI Center before?" Li Bin said. Now, the two sides are sitting at the same purchasing table.

NIO's cost per vehicle increased by approximately 14,000 yuan in the second quarter compared to the end of last year, equivalent to a reduction of about 1.5 billion yuan in one quarter. In the second half of the year, the increase is expected to expand to 16,000 to 17,000 yuan, of which memory accounts for about 10,000 yuan and batteries account for about 3,000 to 4,000 yuan.

Li Bin predicts that this squeeze may continue for another two years, with memory supply not expected to increase significantly until 2028. He also remains cautious about current investments in computing power. If hardware prices drop in two or three years, the replacement cost of equivalent computing power will decrease accordingly, and the early investment may not have been recouped before the equipment is worth its original price.

“If we invest one trillion dollars now, in two or three years, we may only need to spend three hundred billion dollars to buy the same amount of computing power. The extra seven hundred billion dollars may not buy us anything at all.”

In recent years, the number of chips, computing power, and intelligent features have supported the premium pricing of automobiles. Today, the more chips and memory a car has, the more sensitive it becomes to hardware price increases. Features that consumers have become accustomed to are difficult to reduce, and with car prices constantly decreasing due to competition, the profit per vehicle is ultimately squeezed.

Li Bin therefore repeatedly emphasized operating profit. Having money in the company's accounts, whether deposited in banks or invested in wealth management products, can generate returns, but this doesn't prove that the automotive business itself is profitable. His ranking was: "In the past, it was about increasing volume, increasing revenue, and then increasing profit; today, we need to look at the ranking: increasing profit, increasing revenue, and then increasing volume."

NIO's gross margin for vehicles was 18.5% in the second quarter, with the average selling price of the NIO brand at approximately 406,000 yuan, rising further to 434,600 yuan in July. Despite the increased cost per vehicle, the company did not increase deliveries through significant price reductions, and previously offered incentives such as purchase tax guarantees for some models have been withdrawn.

"In a sense, it can be called a price increase, and in fact it is a price increase," Li Bin said.

Qin Lihong also considers used car prices. Frequent price drops for new cars cause existing customers' vehicles to depreciate. When they need to trade in their cars a few years later, they'll have to pay more, and the brand may lose its next order.

As new energy vehicles enter their replacement cycle, resale value will gradually become the focus of competition. For first-time car buyers, price and features are the primary considerations; for second-time buyers, the resale value of their old car will also influence their budget. Securing a sale today through a price reduction might mean missing out on a replacement opportunity a few years later.

The Ledao sits right on this price line. From January to August this year, the average transaction price of the Ledao was 248,000 yuan, and the L60's one-year residual value rate was 74.92%, ranking first among pure electric mid-size SUVs.

Li Bin positions the Ledao at the "high end" of traditional joint venture brands, targeting family users who previously purchased high-spec CR-Vs, Highlanders, and other similar models. "A high-quality family car that's also presentable enough to drive to work," he mentioned several times.

The price of the Ledao will gradually move downwards, but there are no plans to enter the market below 150,000 yuan. "There is no such thing as an exceptionally cheap Ledao," said Li Bin. "We are not magicians, and we cannot really make a high-quality car for 100,000 yuan."

The 150,000 to 200,000 yuan market is large enough, but it's also the easiest to get caught up in the battle over configuration and price. The costs of batteries, intelligent driving, and direct sales channels will not decrease in tandem with vehicle prices. If Ledao relies on low prices to quickly increase sales volume, the group's sales will rise, and NIO's barely maintained profits will be dragged down.

Ledao's share of the group's sales will increase next year. It needs to help NIO cross the 30,000-40,000 unit mark, but it cannot rely on low prices to achieve its goals. Li Bin's standard for Ledao is that each product achieves a leading market share in its respective segment. The L90 is currently ranked first in the 300,000 RMB-level pure electric large SUV market.

Ledao still has over 400 directly operated stores and thousands of sales staff. In third- and fourth-tier cities, stores with all three brands will be added, but sales management and frontline teams will remain separate. R&D has not been merged into NIO, a point Li Bin quickly corrected: "It can't be called a merger; Ledao still has its own independent R&D team. There are many fundamental and platform-based shared aspects that are unified."

"Build strongholds, fight a protracted battle, and take it slow." This is how Li Bin summarizes Ledao's next steps. With Ledao's sales share increasing next year, the first hurdle in this slow battle is profitability.

Dilute past investments

Following the company's reported profitability, R&D expenses quickly became the focus of discussion. In the first half of this year, NIO's sales and administrative expenses were nearly double its R&D expenses, leading the market to question whether the profits were primarily driven by a reduction in R&D spending.

"We've definitely been influenced by some hype," Li Bin said. "We definitely made that money in the first half of the year."

NIO has reduced its focus on application development with lower returns, while retaining its basic research and development in areas such as batteries, cells, chips, operating systems, and autonomous driving.

NIO is using a CBU (Central Business Unit) mechanism to manage projects. Before money is spent, the person in charge must specify who will bear the cost and when the return will be made. If a project is halfway completed and external circumstances change, making the financials unsustainable, it can be stopped. In the past, projects that were driven by annual budgets often continued indefinitely once they were initiated. Now, accounting follows the project's progress.

Li Bin seems to have found enjoyment in it as well. "I am very happy doing business now, I find enjoyment in doing business now," he said. Calculating ROI is not just about filling out a form, but about figuring out all the costs and returns. "This process is just like writing a good article or making a good video, it is very interesting and satisfying."

This method breaks down the problem to individual parts. A car has 17,000 to 18,000 parts, and based on an annual production of 500,000 cars, the number of combinations that need to be processed reaches over 8 billion. Li Bin hopes to use AI to identify cost and efficiency issues.

“Nowadays, large models often have trillions of parameters, so 8 billion is still okay.”

He estimates that it will take about another year to "master" this skill.

In the past, car companies primarily sought economies of scale in their factories. The larger the output and the more fully operational the production line, the lower the manufacturing cost per vehicle. The cost structure of intelligent electric vehicles is more complex, and the software, algorithms, and power supply networks also need to find more users.

The dedicated R&D team for Firefly consists of only a few dozen people. "The efficiency is much higher than everyone imagines," Li Bin said.

These dozens of people don't develop a car from scratch. The battery, electronic and electrical architecture, software, and supply chain come from NIO's existing platform. The team is mainly responsible for product direction and cross-departmental collaboration. According to data from China Automotive Insurance, the Firefly has ranked first in the high-end small car market for 15 consecutive months.

The limited price and profit margins of high-end small cars make it impossible for them to support a complete R&D system independently. The establishment of Firefly is due to NIO's past investments in underlying technologies. On June 18th, NIO simultaneously pushed out system updates to over 700,000 vehicles across two brands and three platforms; the fifth-generation battery swapping stations also began to be compatible with full-size models from the three brands.

Traditional automakers share chassis, engines, and supply chains, while smart electric vehicles add chips, operating systems, algorithms, and charging networks. A single software update can cover hundreds of thousands of vehicles, and a battery swapping station can serve users across different price ranges.

If each brand maintains its own R&D, software, and distribution channels, selling an additional brand simply incurs more costs. The more underlying capabilities NIO, Ledao, and Firefly share, the more likely past investments will be amortized through increased sales volume.

Vehicles already sold are also starting to contribute new revenue. NIO's service and community business revenue exceeded 10 billion yuan last year, and approximately 5.8 billion yuan in the first half of this year. Repair and maintenance, used cars, software subscriptions, and energy services are all growing along with the number of vehicles on the road.

The intelligent assisted driving subscription has already generated a small amount of revenue. There are approximately 40,000 used NIO second- and third-generation models, of which about 8,000 maintain regular subscriptions, generating tens of millions of yuan in related revenue.

"It's only for used car users," Li Bin said. "With tens of thousands of users, we've already generated tens of millions in subscription revenue."

Li Bin views the revenue generated from services and the community as NIO's more realistic second growth curve. The third growth curve may come from energy or physical AI, but that will be further in the future. Currently, cars remain NIO's most important revenue stream. After a car is sold, software, energy, maintenance, used cars, and trade-ins further extend the transaction.

Battery swapping networks have a slower recycling rate. Once a station is built, depreciation and operating costs begin to accrue; only as the number of vehicles served increases will the cost per swap decrease. With three brands sharing the fifth-generation battery swapping station, NIO doesn't have to wait for its main brand's sales to grow; it can also increase the utilization rate of individual stations through Ledao and Firefly.

Li Bin did not set a higher profit margin target for NIO. "Our revenue in the first half of the year was close to 60 billion yuan, and it will definitely exceed 120 billion yuan for the whole year. With such a revenue scale, it is problematic that we have not yet achieved sustained and stable profitability."

New products, basic technologies, and battery swapping still require continued investment. His bottom line is that profits from other businesses should cover these investments, and after that, the company should still be profitable.

In the next financial report, the sales share of Ledao will continue to increase, and the rise in memory prices will be more fully reflected in costs. Whether the money NIO has invested over the past decade can be shared by more models and users will become clearer then.

Risk warning and disclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.