Gasgoo Munich- On November 10, AVATR Vice President Yong Jun took to Weibo to address rumors circulating online about a "600 million yuan junkyard." He clarified that these are not wrecks, but 1,300 test vehicles worth roughly 600 million yuan. These cars have endured extreme heat and cold, collisions, and water immersion. Their frameless doors have been opened and closed countless times. After testing, some are scrapped directly—they simply cannot be sold to customers.
Gao Quanli, deputy director of AVATR's office, dubbed this parking lot a "base for battle-scarred cars." He said these vehicles have "traveled ten thousand miles and processed ten thousand data sets." Having completed their mission and returned to their origin, the scars on their bodies are "not flaws, but medals."
That's 1,300 cars for 600 million yuan—an average of about 460,000 yuan per vehicle. To an outsider, an automaker spending 600 million yuan to build over a thousand non-saleable cars, then clearing out warehouse space to store them, looks like burning cash.
Cars That Finish Testing Shouldn't Enter the Sales Chain
The first barrier to selling these vehicles is regulatory. They haven't been entered into the "Catalogue of Road Motor Vehicle Production Enterprises and Products," meaning they lack certificates of conformity, can't be registered, and aren't covered by the "three guarantees" (warranty policy). Under regulations governing motor vehicle products, test cars, prototypes, and trial-production vehicles are not classified as commercial vehicles. They cannot be sold as new cars or registered; after testing, they must be recycled, dismantled, or scrapped by the manufacturer.
More fundamental than the paperwork, however, is the physical state of the vehicles. Engineering prototypes are in the research and development validation phase. Components may be from early batches, with mixed production dates for parts like glass and tires. Calibration and matching are still being iterated.
After undergoing durability, collision, and wading tests, the performance of some components has degraded or been lost entirely. It is difficult for them to meet the quality and technical standards required of new cars rolling off the assembly line, nor do they possess the consistency needed for mass delivery. Their job is to run through extreme conditions—reinforced corrosion, complex road impacts, wading—on behalf of the mass-production models. Once that's done, they are scrapped or sealed off, never entering any sales channel.
Image source: AVATR
That 600 million yuan falls under R&D and verification investment—money spent that does not result in sellable goods. Only a small minority of prototype cars that underwent non-destructive testing and still meet factory standards could potentially enter circulation. The rest, subjected to destructive testing, are destined only for dismantling or permanent storage.
A representative from AVATR told Gasgoo that the actual investment far exceeds 600 million yuan. That figure represents just the cost of the test car fleet; the unit cost of these vehicles is significantly higher than that of mass-produced commercial cars.
Yong Jun stated that AVATR refuses to build "fast-tracked" cars. If a car is rushed, the risk to family safety increases. There are no shortcuts in automotive R&D not a single cent of testing money can be spared. These 1,300 vehicles will absolutely not flow into the consumer market.
What's Being Compressed Is the Invisible Second Kind of Speed
The backdrop to this response is that trust in the industry's validation processes has already taken a hit.
In October, 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. Environmental adaptation validation must last at least one year, including "two winters and one summer" of testing, and reliability verification mileage must not be lower than 30,000 kilometers. Starting January 1, 2027, vehicles that fail validation standards will not be granted announcement. Where companies once dictated the pace, this document turns cycle time and mileage into hard entry barriers.
Data shows just how much cycles have been compressed. In the internal combustion engine era, a new model typically took over 36 months from drawing board to assembly line. The EV industry has slashed that to an average of 24 months, with some startups bringing cars to market in just 12 to 18 months. In the first half of 2026, 576 new passenger vehicle series were launched in China. Of these, 480 were minor annual updates, meaning truly new models are rare, with most iterations built on existing platforms.
Validation for new energy vehicles is more complex than for internal combustion cars. Beyond traditional components like the body, chassis, and suspension, the battery, motor, and electronic control systems, autonomous driving algorithms, and electronic-electrical architecture must all be tested. The matching between software versions and hardware design also needs to be proven in real-world conditions. Theoretically, the requirements are higher, yet the actual cycle is shorter, concentrating immense pressure on the testing phase.
Image source: AVATR
Summarizing public reports, this compression occurs mainly in two areas. First, design verification and production verification stages are merged, and full-cycle durability testing is cut. Second, there is an over-reliance on simulation and AI testing, reducing the frequency of real-world durability and "three-highs" (extreme temperature, altitude, and humidity) tests. Some real-world tests are conducted merely to generate marketing material.
Consequently, there are two kinds of speed in a car: the visible speed of smoother screens, richer cabin features, and monthly over-the-air updates that users enjoy after purchase; and the invisible speed of cutting corners—running fewer miles than required or skipping seasons. When steel plates will crack under repeated stress depends on material, frequency, and temperature. Tests involving salt spray, humidity, and extreme cold cannot be rushed; results won't appear early just because the schedule is tight.
The boundaries for software and hardware validation are not the same. OTA updates can push software calibrations and functional logic. But hardware—body structure, battery system, chassis, corrosion durability—must be thoroughly validated in the design phase. If it isn't, software updates later on cannot fix it.
Media investigations have uncovered similar cases: consumers complained of rusting vehicles, traced back to insufficient heat and corrosion testing; some brands saw air springs leak and bodies collapse in extreme heat. Other reports noted that durability tests for high-frequency components rely too heavily on computer simulation, with limited hands-on testing.
AVATR stated that,in terms of testing cycles, environmental adaptation validation alone requires "two winters and two summers." The AVATR 07 must also complete over 2 million kilometers of durability testing during the vehicle development phase. Cars from the R&D and manufacturing stages will not be delivered to customers.
By dismantling cars and opening its warehouses, AVATR has effectively put a previously invisible validation process on public display. But live-streamed teardowns and open warehouses can only show production lines, prototypes, and dismantling processes—they cannot replace full-scale verification data. For automakers that haven't opened their test grounds, the outside world has no way to judge how far their validation has gone. Consumers have little information to verify, mostly limited to development cycles and crash test scores.
