Invest1 distinct publisher3 min readUpdated
165 new models in six months, development compressed from about five years to two, and no national rule on how much of the testing can be done in simulation.
The Investor · Invest desk

Compiled by The InvestorSomething wrong?How this is made
Chinese manufacturers put 165 new vehicle models on the market in the first half of this year, roughly double the fewer than 90 launched in the same period of 2016, according to the business outlet Yicai [1][2]. The same reporting says development timelines have fallen from about five years in the combustion era to about two years for EVs, which is roughly the minimum the industry itself assigns to development and verification alone [3][5][6].
That is the whole problem in one line. A half-year is 26 weeks, so 165 launches is more than six a week [7]. The launch rate is not the risk; what was removed to achieve it is. Yicai reports that some manufacturers are shortening timelines by leaning more heavily on laboratory simulation and cutting back road testing with actual vehicles, and that some parts skip separate verification altogether [8]. The industry has coined a term for the output: "rushed cars" [9]. Some makers now claim they can launch three new models a year [4].
The load-bearing detail for anyone with capital in the supply chain is the second half of that sentence. A part destined for a production vehicle is supposed to clear design verification and production verification, and some in the industry warn that completing only part of that sequence, or narrowing its scope, could widen quality variation during vehicle production [11]. Variation is a volume problem rather than a prototype problem: it does not announce itself in one test mule, it distributes itself across units already delivered. That is the shape of exposure that ends up in warranty accruals and field campaigns, and it sits with suppliers as much as with the brand on the badge.
The regulatory floor is lower than the marketing suggests. Combustion vehicles must complete 30,000 kilometres of reliability road testing; new-energy vehicles must complete 15,000, half as much, and China is pushing to raise the EV requirement to 30,000 [13][14]. A research and development official at GAC noted that 30,000 kilometres is not a long test by industry standards, since a vehicle driving 500 kilometres a day covers it in about two months, and that many new models test far beyond the standard yet still develop breakdowns or quality problems [15]. At that pace the current EV floor is about 30 days of driving [16]. Yicai identifies the real gap as the absence of any national standard, beyond minimum mileage, for the ratio of real-vehicle testing to computer simulation, leaving manufacturers wide discretion over methods, the number of vehicles and the number of tests [17]. Even when the mileage box is ticked, thoroughness is hard to judge from outside [18].
The dispute is live inside the industry. Li Fenggang, president of Beijing Hyundai, criticised rivals' development practices at the 2026 China Auto Forum last month, arguing that consumers become "test personnel" who find the defects in use [10]; he also runs a joint venture whose slower product cycle is the one being outrun. Lu Fang, chief executive of Voyah, rejects the framing, saying what matters is not how long development took but whether the verification process is in place and sufficient verification was carried out [12].
Three things to track. Whether the 15,000 to 30,000 kilometre proposal for new-energy vehicles becomes a rule rather than a push [13]. Whether any standard emerges for the simulation-to-road ratio, which is where the discretion currently sits [17]. And whether tier-one suppliers begin insisting on complete design and production verification before shipping, since incomplete validation is the part of this that lands on their balance sheet [11].
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
Bringing a single vehicle to market typically requires dozens of stages including market research, design, technology development, reviews, selection of parts suppliers, simulation and real-world road testing, and the development and verification process alone is known to take at least about two years.
Yicai identified as a problem the absence of clear national standards, beyond minimum mileage, for the ratio between testing with actual vehicles and computer simulation, leaving considerable room for manufacturer discretion over test methods, the number of vehicles and the number of tests.
Even when a standard such as minimum mileage is met, it is difficult to judge how thoroughly the actual verification was carried out.
A research and development official at GAC said a 30,000-kilometre road test is hard to consider excessively long for the auto industry, since driving 500 kilometres a day covers that distance in about two months, and noted that while many new vehicles road-test far beyond the standard, some models still develop breakdowns or quality problems.
China rolled out 165 new vehicle models in the first half of this year, according to Chinese business outlet Yicai on the 20th.
That figure is roughly double the number from the first half of 2016, when new models fell short of 90, a twofold increase over a decade.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Sourced reporting, single outlet, no outcome data
The hard numbers are specific and internally consistent: 165 launches versus fewer than 90 in 2016, five-year to two-year cycles, a two-year verification floor, and 30,000 km versus 15,000 km mileage minimums. But the cluster contains one publisher relaying a second outlet (Yicai), the shortcut claims rest on unnamed industry sources plus one anonymous JV engineer, and no defect, recall or warranty data connects compressed cycles to field failures. The most load-bearing structural finding, the absence of a simulation-ratio standard, is attributed rather than documented against a published standard.
Fast cadence documented; shortcut prevalence not quantified
The launch cadence itself is well documented and clearly widespread: 165 models in six months across the market, with development cycles compressed industry-wide from roughly five years to two. What is only anecdotally adopted is the specific shortcut behaviour: one door-handle example at 80% simulation, one engineer describing single-vehicle part validation, and 'some manufacturers' language throughout. So the enabling condition is broadly adopted while the alleged corner-cutting is evidenced at example scale only.
Safety alarm outruns the outcome evidence
The framing, safety fears and arithmetic that 'does not close', is somewhat ahead of what the sourcing establishes. The structural gap is real and well argued: a two-year cycle against a two-year verification floor, a NEV mileage minimum at half the combustion level, and no rule on simulation share. But no failure-rate, recall or crash evidence links the practice to harm, the shortcut claims are unnamed, the most vivid data point is a single door handle, and the article's own sources push back: the Voyah CEO disputes the premise and a GAC engineer says the mileage bar is not onerous while noting faults occur even in heavily tested cars, which cuts against blaming cycle length specifically.
Competing commercial voices on both sides of the quality charge
Nearly every attributed voice has a stake. The loudest critic is the president of Beijing Hyundai, a foreign joint venture losing share to the domestic entrants whose speed he criticises, and the criticism was delivered at an industry forum. The rebuttal comes from the CEO of Voyah, a Chinese EV brand with a direct interest in defending domestic development practice. The remaining shortcut evidence is anonymous, and the publisher is a Korean business outlet covering the competitive threat to Korean and Japanese automakers, closing on Chinese makers overtaking global sales rankings. The pending 30,000 km standard also creates regulatory incentives for incumbents with slower cycles.
Structural facts solid, causal claim thin
Confidence is moderate-low overall. The countable elements, launch volumes, cycle lengths, and mileage standards, are stated precisely and hang together, so the structural argument that verification time has been squeezed is credible. The step from there to a safety problem is single-sourced, anonymous in its key detail, contested on the record by two named industry engineers and executives, and unsupported by any outcome data. A second independent source or regulator statement would move this materially.
invest
Alo's 10 million yuan opening day in China was a celebrity drop, not a distribution strategy1 distinct publisher
product
Voyah's Passion S starts at 229,900 yuan. That figure, not the tunnel loop, is the input1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
en.sedaily.com
1 article · August 20, 2026