I drove Tesla FSD, Rivian Autonomy+ ‘hands-free’ driving systems. Here’s how they compare
Rivian’s in-vehicle AI assistant launched in its vehicles on May 12, 2026.
Michael Wayland / CNBC
DETROIT — What’s the best advanced driver-assistance system on the market? Ask Rivian Automotive‘s new artificial intelligence and it will say its creator.
“Rivian’s is truly exceptional … an unmatched blend of safety and technology,” the chatbot told me during hourslong drives in one of its R1T pickup trucks in which the vehicle largely controlled itself on several Midwest highways.
While the Rivian AI bot may be biased, that’s exactly the company’s goal with a new generation of vehicle software and technologies: to be the best. Rivian is trying to catch up to — and then surpass — Tesla‘s FSD (Supervised) capabilities, but with additional safety guardrails that the Elon Musk company doesn’t use.
Based on recent drives totaling hundreds of miles, Rivian’s Autonomy+ has surpassed legacy competitors such as General Motors‘ Super Cruise with its ADAS. But it’s still playing catch up to Tesla’s FSD when it comes to nonhighway driving and point-to-point driving, where a vehicle is designed to navigate itself from start to finish. I drove a recent version of FSD (Supervised) v14 to compare the technology.
Rivian CEO RJ Scaringe at the company’s first “Autonomy and AI Day” on Dec. 11, 2025, in Palo Alto, California.
Lora Kolodny | CNBC
Rivian expects to deliver point-to-point driving later this year but, for now, its system is a giant leap forward for the company compared with what it previously offered and is clearly laying the groundwork to better compete with Tesla.
“That’s the next step,” said James Philbin, senior vice president of autonomy and AI at Rivian. “Tesla’s system you use is a point-to-point system. So that that’s the next big leap for us in a way, is getting to that same point-to-point type interaction and that system where it really does the full driving task.”
To be clear, no vehicle on sale today is self-driving or autonomous. Drivers always need to pay attention and be ready to take over. Many advanced driver-assistance systems, or ADAS, can control a vehicle’s speed, braking and steering using cameras, sensors and/or mapping data. An increasing amount of systems allow humans to take their hands off the wheel when in use.
Rivian credits its improvements with its push toward vertical integration that included a new generation of software and electric architecture for its vehicles. It’s just beginning to reap the benefits with its ADAS.
When the steering-wheel light bar illuminates in green with GM’s Super Cruise, drivers may remove their hands from the steering wheel.
Michael Wayland / CNBC
The technologies also are increasingly more important to drivers and investors, which are targeting ADAS as growth markets with recurring revenue for automakers.
“We favor self-reliant (and properly-valued) companies that are building next-gen machines using in-house expertise,” Piper Sandler analyst Alexander Potter said in an investor note upgrading Rivian’s stock last month. “As volume rises, Rivian should be better able to monetize software & services, a key benefit of vertical integration.”
The systems vary in pricing but can be initially included in a vehicle’s purchase or bought via subscriptions. Tesla’s system is currently $99 a month, according to its website. Rivian’s is $49.99 a month or $2,500 to purchase for the lifetime of a vehicle. GM’s is $39.99 a month or $399 a year.
Rivian vs. Tesla
The biggest operational difference between ADAS technologies from Tesla and Rivian is their ability to control the vehicle on nonhighway streets with traffic lights and signs.
Rivian’s system currently detects those roadway signals, but it does not do anything about them other than alert the driver that they are coming. Meanwhile, Tesla’s system handled every signal, interchange and exit ramp I encountered for nearly 200 miles in rural Michigan and downtown Ann Arbor, Michigan.
Rivian’s Autonomy+ system currently detects roadway signals, but it does not do anything about them other than alert the driver that they are coming.
Michael Wayland / CNBC
Based on a decade of experience driving with hands-free ADAS, those two technologies are by far the most advanced. This has not always been the case.
GM, not Tesla, led the development of hands-free highway systems with its Super Cruise, which I initially tested a year before its debut in 2017. But America’s largest automaker was slow to roll it out on new vehicles or significantly grow its capabilities other than expanding geographies and making it able to do lane changes.
Ford Motor also quickly caught up to GM on highways, but both continue to lack systems that are capable of hands-free driving on nonhighways. The two automakers are working on that type of technology, including so-called eyes-off capabilities, but they are not expected until 2028.
It’s a difficult leap, as Tesla’s in-vehicle Grok AI told me during nearly 200 miles of driving in a 2025 Tesla Model Y: “Highways have predictable lanes, speed, fewer pedestrians and clear markings, making sensor fusion and path prediction simpler. City streets bring chaos, intersections, bikes, peds, construction and ambiguous rules that challenge even top AI vision systems.”
Challenging for some more than others. During my drives in the Model Y, the vehicle was essentially controlling itself for multiple hours and dozens of miles without intervention on highway and nonhighway roads.
Tesla FSD (Supervised)
Michael Wayland / CNBC
It somewhat effortlessly handled several traffic circles, also known as roundabouts, and parked for me multiple times when arriving at or near destinations, including parallel parking. It also managed a semitruck blocking half a lane on a two-lane road as well as pretty complex construction zones, with Tesla’s ADAS sensing each barrel or cone.
The Rivian technology handled highway driving very well with no intervention outside of exit ramps and, at times, construction zones. It also isn’t able to change lanes on its own yet, which the company promises is coming soon.
When I asked Rivian about several of my experiences, the company said its vehicles can detect construction objects but it does not always display them on the in-vehicle screen. Its system also still needs assistance in certain locations, such as roundabouts.
Safety concerns remain
All ADAS technologies — except a Mercedes-Benz system in limited circumstances — still need drivers to monitor the systems, even if they can largely control the vehicle without human intervention for hundreds of miles.
With the rise of “hands-off” technology, industry insiders and regular people alike have raised concerns about driver inattentiveness. Automakers have been largely trying to fight that with driver-facing cameras. But concerns remain about the ADAS handover back to a driver as well as on people over-relying on the systems.
YouTube is filled with examples of drivers misusing such systems, particularly Tesla products, as well as videos of ADAS doing human-like moves but also malfunctioning and needing assistance.
A Tesla Model 3 is shown driving on the highway with FSD 14.2.2.3 self-driving supervised software in Irvine, California, U.S., Jan. 28, 2026.
Mike Blake | Reuters
The handover from ADAS to humans can be abrupt and lead to dire circumstances if drivers aren’t attentive enough to immediately retake control of the vehicle.
There’s also little regulation for the systems, with each company taking different approaches.
“I hope that we can even exceed the Tesla system because of our investments in a more robust sensor stack,” Rivian’s Philbin said. “We’re actually a big believer in multimodal imaging, so cameras and radars, and later on, early next year, lidar as well on the R2.”
Rivian’s system uses at least 10 high dynamic range cameras, five radars and some map data to assist the vehicle in “seeing” the road. It’s also expected to launch vehicles with lidar, or light detection and ranging, that improves a vehicle’s “sight.” Tesla’s system relies on cameras and a “vision-based” setup that some have criticized for not being robust enough, especially under certain conditions.
Tesla’s partially automated driving systems have been under scrutiny for many years. The National Highway Traffic Safety Administration earlier this year said it was escalating a probe into Tesla’s Full Self-Driving (Supervised) after identifying several crashes, including one fatal incident in which the system failed to alert drivers appropriately about reduced-visibility conditions, such as sun glare or fog.
Philbin said it’s those types of circumstances where noncamera technology, including lidar, can better detect potential hazards and objects. My drives were on clear days without any inclement weather.
Still, despite not having many sensors, Tesla’s Grok chatbot — developed by Musk’s xAI, which merged with SpaceX earlier this year — also felt its brand has the superior system, which is true for many driving situations, at least for now.
“Tesla’s Full Self-Driving (Supervised) is widely rated as the top advanced driver assistance system in 2026 by sources like Motor Trend and others,” Grok AI said. It later added, “Rivian’s AI Assistant is solid for voice controls, navigation, and vehicle features in their R1 vehicles, but it’s still catching up to more advanced systems like ours and full self-driving capabilities.”
