Toyota bZ: the biggest carmaker says no
Toyota will build a car that can drive itself and sell it as one that cannot, because the driver keeps the liability. The bZ — cheap, competent, 314 miles, hands-on only — is what that decision looks like in a showroom.
The world's largest carmaker has decided not to take the wheel, and it has written down why. In August, the head of Toyota's software division set out a plan for an AI-based driving system arriving in 2028 — and then said the quiet part aloud: "We'll develop Level 4 technology from a technical standpoint, but offer it commercially as Level 2++." Toyota will build a car that can drive itself and sell it as a car that cannot, because the driver keeps the liability. Every other statement in this review follows from that one.
The car in front of you is the evidence. The 2027 bZ arrives in American showrooms this month at $34,980 — six thousand dollars cheaper than the model it replaces, with a native Tesla plug, up to 314 miles of range and a 74.7 kWh pack. It is a thoroughly decent, sharply priced electric crossover, and Toyota is finally selling them: the brand was America's fourth-largest EV seller in the first half of 2026, up 136 per cent. What it does not have is anything an autonomy site would call news. It ships with Toyota Safety Sense 3.0 — hands-on lane centring and adaptive cruise. Not even the newer 4.0, which went to the RAV4 first, and which is itself a sensor-and-software polish rather than a new tier.
Hands-free systems Toyota or Lexus currently sell in the United States. Teammate, the lidar-equipped Level 2 system launched in 2021, died with the Lexus LS after the 2026 model year. A $108,550 Lexus LX has no equivalent of Super Cruise at any price.
The autonomy
The strange thing is that Toyota has all the pieces. It shipped a forward-facing lidar in 2021, years before most rivals thought of it. It has the NVIDIA compute deal, the Arene software platform now live on the RAV4, a $10-billion-ish prototype city at the foot of Mount Fuji, a Level 4 robotaxi partner in Pony.ai, an exploratory arrangement with Waymo that has sat exploratory for seventeen months, and eleven million cars a year of training data. It has pointed almost all of it at manufacturing efficiency and hybrid margin. The one place Toyota does sell genuine urban autonomy to customers is China, where the bZ7 carries a roof lidar and Momenta's navigate-on-autopilot for about $21,000 — technology Toyota buys rather than builds, in a market where it had no choice but to compete on it. Akio Toyoda set the posture in 2021, after an e-Palette shuttle struck a visually impaired judoka in the Paralympic Village at walking pace with a human on the joystick: autonomous vehicles, he said, "are not yet realistic for normal roads." Five years on, that is still company policy, now with a delivery date attached.
"Toyota is not behind by accident. It has looked at eyes-off, priced the liability, and decided the driver can keep it. That is a strategy, not a stumble — and it may still be wrong."
What we're watching
- Whether the Waymo agreement ever becomes more than "preliminary" — Waymo is already testing robotaxis in Tokyo without Toyota.
- The e-Palette shuttle reaching Level 4 in the 2027 financial year — Toyota's first real driverless product, and it isn't a car.
- Whether "Level 2++" survives contact with Europe's revised product-liability rules, which treat software as a product whatever the marketing calls it.
Spec sheet — Toyota bZ (2027, US)
The ai.cars verdict
- A genuinely good, genuinely cheap electric crossover — and Toyota is now selling them in real numbers.
- On autonomy it offers hands-on Level 2 and nothing else; Toyota sells no hands-free system in America at any price.
- The 2028 plan is Level 4 capability sold as Level 2++ so the driver keeps the liability — a deliberate, documented choice.
- Toyota ships lidar-grade urban assistance in China, bought from Momenta and Huawei, because there it had to.
- Verdict: buy it as an EV, not as an AI car. The world's biggest carmaker has decided autonomy is someone else's risk.
How does it stack up? Compare the bZ with the Tesla Model Y, or build your own comparison.
