Rivian told buyers the R2 would be roughly as efficient as a Model Y. Travis Ketchum has now had one for three weeks, and he says he is averaging 2.76 miles per kilowatt hour against the 3.48 or so he would need to reach the rated range on his all-terrain tyres. That is about 21 percent short. He is not softening it either: he says that given the comparison Rivian drew, the company deserves the criticism it has taken. The car is not broken. The number simply is not the number.
In the same week, a car went the other way. Dirty Tesla ran a Model Y L at a steady 75 mph on Michigan highways and logged 222 watt hours per mile over the measured stretch, which he calculates as roughly 370 miles on a full pack. Folding in the battery preconditioning before a supercharger stop, he settles on about 228 watt hours per mile at a 70 mph average, or somewhere around 360 miles. The advertised figure is about 320. He was on the larger 20-inch wheels, and he expects the 19s to do better still.
What separates those two outcomes is not engineering prestige. It is whether the company published a figure it could survive being checked against. Both cars will now be measured constantly, by owners with trip computers and by reviewers with a route and a stopwatch, and the gap between the brochure and the spreadsheet has become the most repeatable story in this segment.
The rest of today is the part no spreadsheet catches. Out of Spec Reviews spent a first ride in Tesla's Cyber Cab mostly cataloguing things you can only find by sitting in one: doors that swing into the pavement, a suspension the host thinks is less comfortable than a Model Y's, no way at all to see behind you. Autogefühl found a Mazda with genuinely good measured efficiency and a lane-keeping system aggressive enough that the first thing they did was switch it off. Neither of those appears on a spec sheet, and both would shape a year of ownership.
What to watch is where the measuring gets done next. CBS News was let inside a Geely plant building Zeekr cars, and its manufacturing vice president made a claim on camera that Chinese-built quality has passed the American-built equivalent. That is his assertion, not a tested finding, and the sensible response is the same one the R2 is getting: wait for somebody outside the company to put instruments on it. Under the deal Canada signed in January, 49,000 Chinese EVs a year now enter at 6 percent duty rather than 100. North American reviewers are about to get their hands on the cars, and that argument stops being rhetorical.
A First Ride in the Cyber Cab, and the Doors Are Already a Problem
Out of Spec Reviews got early access to Tesla's Cyber Cab in Austin, and the first thing Kyle Conner reports is nearly walking into one of its doors on the pavement. He says a door opened into him more than once, and that the presentation framed the swing-out design partly as an accessibility feature for wheelchair users. Cameras stop the door when somebody is close, and there is a physical button and a manual release inside.
Inside, he is impressed by the isolation. He notes it is single-pane glass with a thick seal and calls the cabin serene, with a large, snappy screen and colour-matched lighting that identifies your car in the app. He says the car uses electronic actuators on each of the four calipers rather than hydraulic brake fluid, which he believes is a first for a production car, and that the boot will take 220 pounds.
The criticisms are ride and visibility. He describes audible front motor whine, some in-lane wobble, and a suspension he thinks is less comfortable than a Model Y's, arguing Tesla could have gone softer given the vehicle's purpose. There are no mirrors and no rear camera view available to riders, which he says he would want at night. His verdict: a clever exercise in cheap, durable, high-volume design that does not yet do anything a Model Y on the same network was not already doing.
The Model Y L Beat Its Own Highway Rating
Dirty Tesla took a Model Y L out on the New Jersey-style straight stuff of Michigan highways for a constant 75 mph range test, then ran it down to 4 percent and plugged in. Conditions were close to ideal at 68°F, on the upgraded 20-inch wheels with aero caps still fitted, and he left the climate control on to keep the test realistic.
Over the measured highway stretch he records 50.7 miles in 42 minutes at a 72.4 mph average and 222 watt hours per mile, which he calculates as about 370 miles for a full pack. Once the car began preconditioning for the supercharger, consumption climbed sharply; folding that in, he settles on roughly 228 watt hours per mile at a 70 mph average, or about 360 miles. Tesla's advertised figure is around 320. He notes Edmunds recorded 358 miles on the same 20-inch tyres, but at a roughly 40 mph average, which is a much easier test.
Charging was less remarkable. Plugged in at 4 percent at a 325 kW stall in Brighton, Michigan, he watched it ramp through 171, 179 and 206 kW before settling at 251 to 252 kW, where it held for a few minutes before tapering. He describes the curve as similar to the existing Model Y: strong at the start, then dropping off faster than he would like. He expects the smaller 19-inch wheels to improve the range figure further.
Inside the Geely Plant, a Claim CBS Was Invited to Hear
CBS News foreign correspondent Anna Coren was given access to a Geely plant producing its luxury Zeekr brand, and describes a line running at what she is told is 99 percent automation, with robotic arms working in unison under AI coordination.
The plant's vice president of manufacturing, Zhao Chunlin, who previously worked for GM, told her that Chinese EVs are better than American ones, and went further, saying that even Teslas built in China are better built than Teslas built in America. That is his claim, made on camera at his own factory, and CBS presents it as such rather than as a tested finding. He also said he hopes one day to sell and possibly manufacture the group's cars in the United States, and that fears of Chinese dominance of the global car industry are unnecessary.
The commercial context is the trade deal Canada signed with China in January, which the report says allows 49,000 Chinese EVs a year with tariffs falling from 100 percent to 6 percent. Coren was allowed to test drive the Zeekr 9X, described as a top-of-range hybrid capable of 745 miles on a combined fill and charge, retailing around $70,000, which the report puts at roughly half the price of an Escalade or Tesla equivalent. Exports to Europe and the Middle East begin this month. She also filmed it parking itself with nobody inside.
Three Weeks in an R2, Twenty-One Percent Below the Rated Number
Three weeks into R2 Performance ownership, Travis Ketchum's headline finding is efficiency. He says Rivian's own app shows him averaging 2.76 miles per kilowatt hour against the roughly 3.48 he would need to hit the rated range on his 20-inch all-terrains, about 21 percent short. For context he gives his R1S numbers too: 2.26 against a required 2.64, roughly 15 percent short, which works out to about 325 real miles. With only an 88 kilowatt hour pack, he argues, the shortfall hurts more. Given Rivian's claim that the R2 is as efficient as a Model Y, he says the company deserves the criticism it has received.
The thing bothering him daily is smaller. He describes an intermittent rattle or flutter from the HVAC that appears under certain airflow conditions in auto mode and that he can reproduce reliably. He says service technicians heard it, agreed it sounded odd, investigated, and returned a determination of operating as expected. He also says dozens of other early owners replied to his social posts describing the same noise.
He is critical of three finishes: pillars in stiff textured plastic rather than headliner textile, easily scuffed plastic on the lower doors, and the same plastic through the cargo area. He rates the 20-inch all-terrain upgrade as worth its roughly 7 percent range cost, and expects Premium to be the sweet spot once the launch bundle ends. He says he would buy it again.
The Mazda That Measures Well and Beeps Too Much
Autogefühl's Thomas drove the Mazda CX-6e, an electric SUV built on a Changan platform and adapted for European roads, priced from around €50,000 in the Takumi trim he deliberately chose as the least expensive. It is rear-wheel drive only, quoted at 7.9 seconds to 100 km/h on 19-inch wheels or 8.1 on the optional 21s fitted to his car.
The efficiency surprised him. At a mild 19°C with topographic equalisation applied, he recorded 16 kilowatt hours per 100 kilometres at 100 km/h, which is 3.8 miles per kilowatt hour and works out to roughly 460 kilometres or 290 miles of real-world range. At 130 km/h that fell to 19 kilowatt hours per 100 kilometres, or about 390 kilometres. He notes this was on the larger wheels, so the 19s should do better, and that colleagues recorded slightly worse figures. He puts the battery at 74 kilowatt hours net against a quoted 78 gross, and says it took 130 kW at a 70 percent state of charge, unlike the Mazda6e sedan, which he says had charging problems and is now getting the same updated LFP pack.
His complaints are software. The lane-keeping assist he calls far too aggressive and had switched off within minutes; the adaptive cruise sits too close to the adjacent lane; and there are, he says, too many warnings and chimes. The seats, the fabric-covered surfaces and the frunk all draw praise. The turning circle he measures at 12.4 metres, which he compares to a truck's.
Bottom line: Publish a number you can survive being checked against. The Model Y L cleared its rating by a wide margin because Tesla set the rating low; the R2 is taking a beating because Rivian invited a comparison with the most efficient car in the class and then did not meet it. Both cars are good. Only one of them has a credibility problem, and it was created by a marketing line rather than by an engineer. If you are ordering an R2, order the wheels you want and budget the range you actually see, not the one on the configurator.
Commentary on third-party videos. Figures and claims are as presented in the sources and have not been independently verified. Spotted an error? Tell us and we will correct it.