This week in car rental & carsharing

Hertz is having a rough time selling off its used cars, which is a problem when your whole recovery plan involves selling off a lot of used cars. The used-car market softened, the EVs they famously bulk-ordered are worth less than expected, and now the guidance looks worse than it did a quarter ago. When your exit strategy depends on resale values, timing matters more than the strategy itself.

Polestar is Swedish-headquartered, makes EVs in China, and apparently that's now enough to get quietly locked out of the US market from 2027. Not tariffs this time — a separate "connected vehicle" rule that treats Chinese-manufactured cars as potential data security risks. The argument being that a car with GPS and driver assistance software is basically a surveillance device. Maybe. Or maybe it's just a useful way to freeze out competitors without calling it protectionism.

Either way, Polestar already sells 94% of its cars outside the US, so the actual damage is limited. Europe will survive the extra attention.

Lithium prices in China apparently jumped 190% since last summer, and CATL noticed. So now they're pushing sodium-ion into actual cars — somewhere between 10,000 and 20,000 EVs this year, plus a stationary storage system unveiled in Munich. The CEO is talking 100 GWh annual sodium capacity within a few years.

Whether "we're done depending on lithium" lands as a strategy or a negotiating position probably depends on how the next few quarters go.

Dallas is up 275% and everyone's acting like that's a clean data point. It's not — small base numbers do that. Still, when Argentina, England, and the Netherlands all land in your city in the same week, apparently people rent cars for longer and want to drop them off somewhere else entirely. Which makes sense if you're following a team across three countries.

The one-way and longer-duration numbers are the quietly interesting part. People aren't just showing up for a match — they're building trips around it.

Hertz is blaming "softness" in the used car market for disappointing numbers, which is a polite way of saying the cars they bought at peak prices are now selling for noticeably less. They over-fleet, residuals drop, the math stops working. This has been visible for a while. The EV write-down drama wasn't that long ago. At some point "market conditions" starts to sound like a recurring excuse rather than bad luck.

768,000 shared cars by 2030 sounds impressive until you remember how many of these projections quietly get revised after the deadline passes and nobody's looking.

The carsharing sector has a long history of confident numbers that aged badly — services folded, fleets shrank, cities turned hostile. Maybe this time the math works out. Or maybe 2030 is just far enough away to put a big figure on a slide without anyone asking too many questions yet.

Tesla's VP of AI goes on X to say the driver floored it manually. The family says Autopilot was on. NHTSA and NTSB both opened investigations. Tesla also just quietly settled a different fatal FSD crash from 2023.

This is the moment the "AI and robotics company" branding gets stress-tested in actual courtrooms and federal offices, not keynotes. FSD is the product that funds the whole narrative. When it crashes, the scrutiny lands somewhere more serious than a comment section.

Amazon's robotaxi subsidiary is about to start proper production — a 20,000 m² factory in Hayward, California, ramping toward 100 vehicles a week eventually. The car itself got a refresh: greener seats, repositioned reflectors that change colour, better cup holders. You know, the stuff that matters when there's no driver to blame for anything.

The part worth watching is the "subject to regulatory approval" attached to basically every sentence. Great interior design. Pending permission to exist.

Hertz stock is down again, which at this point feels less like news and more like a recurring calendar event. The company has been quietly accumulating a mix of debt, unsold EVs, and operational chaos for long enough that a red day on the ticker barely needs an explanation anymore. Whatever the specific trigger today, the underlying story hasn't really changed in a while.

Nissan, ten academic partners, a £10 million government cheque, and three years of research to figure out whether a car can send electricity back to the grid. Which, to be fair, is a real question worth asking. Just funny that the Leaf — a car that's been around long enough to feel nostalgic — is the test platform for technology that's supposed to feel futuristic.

V2G has been "almost ready" for a while now. Maybe this time the grid will actually notice.

768,000 shared cars by 2030, 141 million users — someone did the math and it pencils out, apparently. The less obvious part of the forecast is that none of this actually works without a fairly unglamorous layer of telematics, remote access, billing software, and fleet dashboards running underneath. The car is almost the easy part.

Free-floating services especially — thousands of vehicles scattered across a city, unattended, needing to be found, unlocked, billed, and repositioned — are basically a distributed IoT operation wearing a mobility brand. Which means the interesting business might not be the cars at all.