Before a new car hits the market, it often goes through a long, arduous process that few ever see. It might circle test tracks endlessly, undergo rigorous lab testing, or even be crashed and dismantled—never making it to the sales floor.
These are known as test vehicles.
Recently, an article published by Gasgoo titled Why 600 Million Yuan Worth of Battle-Damaged Cars Can' t Be Sold to Users went viral. Within a week, it had racked up over 200,000 reads and more than 200 comments.
Why has such a routine industry practice—crash testing—sparked such intense debate?
Where Do the 'Test Cars' Go?
This fleet comes from a parking lot revealed by Avatr. Yong Jun, a vice president at the company, once described this 600-million-yuan graveyard: it holds 1,300 test vehicles that the company values at roughly 600 million yuan.
These vehicles have survived extreme cold and heat, been crashed, submerged, and had their frameless doors opened and closed thousands of times. Some are scrapped outright after testing and can never be sold to customers. Gao Quanli, deputy director of Avatr' s office, calls the site the Battle-Damaged Car Base.
The comment section exploded after the article ran. As readers debated whether these cars could be sold, industry insiders weighed in. One commenter using the alias Xiao Hua noted that prototype and test vehicles nearing the production stage often have necessary mass-production parts installed and the latest software flashed before being sold to employees at a discount. Earlier-stage vehicles are typically scrapped directly. “This kind of disposal is not uncommon within the industry,” the commenter wrote.

Image source: Avatr
Industry insiders further clarified the boundaries for internal circulation. Only pre-production vehicles that have reached the PT (Production Trial) stage, underwent only non-destructive testing, and passed inspection to meet factory standards can be refurbished internally and sold to employees at a discount.
However, mule cars and early engineering prototypes from before the PT stage, along with vehicles that have undergone destructive testing like crashes, are in principle prohibited from external sale. Once testing is complete, most are dismantled by qualified recyclers commissioned by the automaker. A small number, subject to internal approval and strict record-keeping, may be sealed within the factory for technical review, internal training, or display—though they are strictly barred from the civilian registration market.
Another practitioner mentioned that their company sells some test vehicles near mass-production status to employees at a discount. Refurbished, these cars still hold value and are often sought-after internally. That' s because test vehicles typically use precision-machined parts worth “dozens of times more” than their mass-production counterparts.
One of the most hotly debated topics in the comments was the cost breakdown. With 1,300 battle-damaged cars valued at 600 million yuan, the cost per vehicle comes to roughly 460,000 yuan.
One commenter, using the alias Xiao Ming, pointed to the electromechanical industry, noting that because prototypes are custom-made in a new state, costs can run several times higher than mass-produced units. Another netizen highlighted PCB procurement costs: “When you buy 10, the unit price is 5,000 yuan; when you buy 10,000, it drops to 5 yuan. Early R&D lacks volume, so costs are naturally higher” (Netizen comment for reference only). Additionally, the cost of body-in-white shells used in trial production is steep.
For brands with broader lineups, the scale of their test fleets is likely astronomical. An intelligent driving R&D professional broke down the math in the comments: even for a model selling just 1,500 units a month, conducting driving, parking, and active safety tests across 50 representative cities—deploying at least 10 cars per city and 50-plus in Tier-1 cities—would require a fleet of over 500 test vehicles for this single test category.
There is a widely cited rule of thumb in vehicle development known as “5+50+200,” which summarizes the rollout rhythm of prototype vehicles: as development progresses, the number of test prototypes increases step by step. For mainstream automakers juggling multiple projects, cumulatively investing thousands of test vehicles is standard practice.
The cost of a test prototype cannot be directly compared to the price of a mass-produced car; costs vary wildly depending on the development stage. Total validation investment also covers small-batch custom parts, sensors, data acquisition systems, vehicle modifications, testing operations, and calibration labor.

Image source: Avatr
Avatr also confirmed to Gasgoo that the actual investment in these damaged vehicles far exceeds 600 million yuan. The display only accounted for the cost of the parking lot; the cost per test car is significantly higher than that of a mass-produced model.
Moving a new car from blueprint to assembly line involves multiple development stages. Early on, automakers build engineering prototypes to validate design schemes and component matching, while parts and software are still in flux. As development advances, the car approaches mass-production status, and manufacturers switch to parts and processes closer to final specifications to check vehicle performance, durability, and production consistency.
One bottom line is clear: test vehicles, prototypes, and trial-production cars are not commercial products and do not appear in the Road Motor Vehicle Manufacturers and Products Announcement. If the vehicle isn't listed in this catalog, it cannot be issued a factory certificate of conformity, cannot be registered for license plates, and is not protected by the “three guarantees” (warranty) rules. Strictly disguising them as new cars for sale is prohibited.
The Supreme People' s Procuratorate has also issued clear guidelines: assembling and refurbishing scrapped vehicles for registration, or selling unqualified vehicles under the guise of low-priced “inventory cars,” can lead to criminal charges for producing or selling counterfeit or substandard products if the circumstances warrant. The refurbishment and market release of non-compliant test prototypes fall into this same high-risk category.
A constraint even greater than regulatory certification is the condition of the vehicle itself. The mission of an engineering prototype is to withstand extreme conditions on behalf of the mass-produced model.
After enduring durability cycles, crashes, and wading tests, many components suffer performance degradation or failure, failing to meet the quality standards of new mass-produced cars—let alone the consistency required for delivery. Even if some prototypes that underwent non-destructive testing have intact hardware, their non-standard, trial nature ensures they never enter the sales chain.
Why Did Routine Testing Go Viral?
Within the automotive industry, deploying hundreds or thousands of test vehicles for extreme cold and heat tests, as well as crash and durability trials, is considered “basic skills” for every compliant automaker.
As one comment noted: “This is standard procedure for car manufacturers. Many top-tier firms spend even more.”
Yet a curious phenomenon emerged: this industry “standard,” once pushed into the public eye, ignited over 200,000 reads and hundreds of fierce debates. Public opinion split into two camps. On one side, there was high-praise agreement: “Safety trust is bought with money,” and “Only safety backed by real silver and gold is reassuring.”

Image source: Gasgoo
On the other side were vigilant questions: “Has verification testing actually decreased?” “Will these costs eventually be passed entirely to car owners, even if the cars they buy are still full of bugs?” Some netizens bluntly expressed concern that certain automakers are “selling test cars or early problem models to first-batch users,” effectively “making owners the guinea pigs.”
The fact that routine testing struck a raw nerve essentially reflects a widespread lack of consumer trust in vehicle quality and safety in the era of smart EVs.
In recent years, competition in the new energy vehicle market has become white-hot, and the iteration speed of new models has constantly been broken. When “18 months to complete vehicle development” and “a model refresh every six months” become marketing tags for some companies, consumers—while enjoying advanced intelligent driving and flashy cabins—are left with a nagging question: with car-making moving this fast, have those invisible, basic verifications really been done thoroughly?
This anxiety is not unfounded. In the era of smart vehicles, the frequency of recalls hasn't dropped—it has risen, and the reasons are increasingly varied.
In August of this year, the State Administration for Market Regulation announced recalls by several automakers, including Xiaomi, Tesla, Leapmotor, and XPENG. The issue pointed directly to life-saving escape routes after severe collisions. The emergency mechanical door handles inside the affected vehicles were too close in color to the interior trim, making them difficult to identify and operate in an emergency. If a severe crash caused the low-voltage system to fail, it could directly hinder occupants from opening doors to escape.
The issue behind this recall does not involve cutting-edge intelligent technology, but rather the fundamental bottom line of car manufacturing.
Beyond recalls, the industry has recently seen a spate of basic quality defects that “should never have happened” under traditional car-making logic.
During repeated emergency braking tests at a closed site, the brake pedal bracket on a new model from one brand snapped, disabling the foot brake. In other cases, owners complained that front and rear wheels were mismatched; the company traced the error to a coding mistake during sub-assembly and offered free inspections and replacements. When basic safeguards fail, consumers inevitably ask: have the safety and consistency of the cars delivered to them been properly vetted?
When the market is plagued by “rookie mistakes,” the resonance caused by automakers showing off their “battle-damaged car bases” stems from a simple truth: consumers are desperate to see “real testing.”
“Slow Car-Making” Can' t Just Be a Slogan
The noise in the public sphere is forcing the entire industry to re-examine the balance between speed and quality.
Improving R&D efficiency is not inherently a sin. Gao Hua, a third-party technical researcher, told Gasgoo that the drastic shortening of vehicle development cycles is due, in part, to the platformization and maturing of core technologies like batteries, motors, and electronics, and in part to the extensive use of lab simulation in early development testing.
Virtual verification, bench testing, and digital tools can indeed replace some repetitive tests, cutting costs and boosting efficiency.
But no matter how powerful simulation tools become, they cannot replace the tempering of the real physical world. Hu Sheng, a researcher at one brand, used the movie The Truman Show to make a precise analogy: “Can a Truman who grew up inside the show be the same as a Truman who grew up in the real world?”
A virtual environment can simulate preset scenarios, but it struggles to fully replicate the complex and ever-evolving physical wear and tear of real-world roads.
Metal fatigue, the cycle life of battery packs after thousands of charge-discharge cycles, and the aging of chassis bushings all require real vehicles to be driven in extreme environments—high heat, humidity, cold, and high altitude. Intelligent driving systems cannot rely solely on closed tracks or simulations. Faced with sudden cut-ins, construction detours, and extreme weather on real roads, the system must not only complete preset tasks but also leave a safe margin for driver intervention when pushed beyond its limits.

Image source: Great Wall Motor
The boundary between fast and slow should be determined by technical risk and safety baselines, not by marketing milestones or launch schedules. Faced with quality concerns exposed during the industry's rapid expansion, rigid regulatory constraints have tightened across the board.
In August 2026, the Ministry of Industry and Information Technology, the Ministry of Public Security, the Ministry of Ecology and Environment, and the State Administration for Market Regulation jointly issued a notice to launch a one-year “special campaign on production consistency and quality improvement for road motor vehicle products.” The campaign targets four core areas: production consistency, reliability, durability, and new technology verification, requiring companies to conduct self-inspections and corrections.
Notably, “new technology verification” has been singled out as a regulatory priority. The notice explicitly requires companies to assess the impact of innovative designs on user habits, risks of misuse, and extreme usage scenarios (such as hidden door handles, screen-based gear shifting, and emergency mechanical handles). It also imposes stringent requirements for the testing and verification of combined driving assistance and autonomous driving functions.
This implies that regulation has shifted from static “pass-fail” standards to a penetrating examination of potential risks in real-world usage scenarios. A deeper challenge lies in the coordination between automakers and the supply chain. “Slow car-making” is not a solo act by the OEM; it is a full-chain engineering project spanning component design, supplier manufacturing processes, and the consistency of vehicles rolling off the assembly line.
A supply chain insider admitted to Gasgoo that in recent years, OEMs have aggressively compressed development cycles and squeezed procurement costs to the limit. This has left some parts suppliers without enough time or profit margin to conduct adequate mold verification and durability testing. “Samples are qualified in the lab, but once they hit mass production, batch consistency issues surface.”
This special campaign by the four departments explicitly included supplier management, procurement standards, incoming inspection, and production processes in its scope—hitting this vulnerability right on the head.
If automakers truly want to hold the quality line, they must front-load verification standards and reasonable timelines into the supply chain, giving parts manufacturers the necessary space for R&D and testing. At the same time, they must establish an internal “veto” mechanism—never cutting critical road tests or shortening verification mileage just to rush a new car to market.
Summary:
600 million yuan worth of “thousands of battle-damaged cars” cannot directly prove the quality of a single model, nor does it represent an entire company's R&D level. Ultimately, consumers want to see that problems found in testing are resolved, that mass-produced cars maintain stable quality, and that companies respond promptly and take responsibility when defects arise.
Because car quality is never decided by a single test before launch. It runs through the entire process of R&D, parts procurement, manufacturing, and after-sales use. Automakers can speed up R&D, but they cannot offload necessary verification onto consumers.
So-called “slow car-making” must ultimately be implemented in every test, every component, and every car delivered to the user.









