General Motors has spent over a decade and billions of dollars pursuing fully autonomous vehicles through its Cruise subsidiary. After a high-profile launch in San Francisco, a serious pedestrian incident in late 2023, and a subsequent operational reset, the trajectory of GM's self-driving car ambitions has shifted significantly. Here's what buyers, riders, and industry watchers need to know about where the technology actually stands.
SAE Automation Levels: Where GM's Robotaxi Sits
Before evaluating any autonomous vehicle, it's essential to understand the SAE International framework that defines what "self-driving" actually means. Consumer vehicles like the Tesla Model Y with FSD (Supervised) or GM's own Super Cruise are Level 2 systems — they require continuous driver attention. Robotaxis aim for Level 4, meaning the vehicle handles all driving tasks within a defined operational domain (geofenced area, specific conditions) with no human behind the wheel.
| Level | Name | Who Drives? | Example |
|---|---|---|---|
| Level 0 | No Automation | Human does everything | Basic cruise control |
| Level 1 | Driver Assistance | Human drives, system assists one function | Adaptive cruise OR lane-keep |
| Level 2 | Partial Automation | Human drives, system assists steering + speed | Super Cruise, Autopilot, BlueCruise |
| Level 3 | Conditional Automation | System drives in limited scenarios, human on standby | Mercedes Drive Pilot (selected markets) |
| Level 4 | High Automation | System drives fully within geofence, no human needed | Waymo One, Cruise (pre-2023 SF) |
| Level 5 | Full Automation | System drives everywhere, all conditions | Does not exist commercially |
Cruise's driverless service in San Francisco operated at SAE Level 4 — no safety driver, no steering wheel in some vehicle generations. GM's consumer-facing Super Cruise and Ultra Cruise remain Level 2 and require the driver to monitor the road at all times. These are fundamentally different products serving different use cases.
Where Can I Use GM's Robotaxi Service Right Now?
Current status (2026): Cruise's commercial driverless ride-hail service is not currently operating in any city as a paid public service. Following the October 2023 San Francisco incident — in which a Cruise vehicle struck and then dragged a pedestrian who had been thrown into its path by another car — the California DMV suspended Cruise's deployment permit. GM subsequently paused all driverless operations nationwide and initiated a phased, supervised return-to-road process.
As of 2026, GM has been gradually reintroducing Cruise-origin vehicles with human safety drivers in select test markets to rebuild operational data and regulatory trust. Fully driverless commercial service has not been re-approved in any jurisdiction. Here's the operational timeline and current status:
| City | Peak Service (2023) | Current Status (2026) |
|---|---|---|
| San Francisco, CA | 24/7 driverless commercial service | Suspended; supervised testing only |
| Phoenix, AZ | Limited driverless service | Suspended; supervised testing phases |
| Austin, TX | Testing with safety drivers | Supervised testing |
| Houston, TX | Announced expansion | Never launched commercially |
| Dallas, TX | Announced expansion | Never launched commercially |
For riders seeking an actual Level 4 robotaxi experience today, Waymo One remains the only broadly available commercial driverless service in the U.S., operating in Phoenix, San Francisco, Los Angeles, and select other markets.
GM's Autonomous Tech Stack: Sensors, Software, and the Origin Vehicle
Cruise took a sensor-fusion approach fundamentally different from Tesla's camera-only strategy. The technology stack on the Cruise vehicle (based on the Chevrolet Bolt EV platform, with the purpose-built Cruise Origin as the next generation) includes:
- LiDAR: Multiple roof-mounted and body-integrated units providing 360° 3D point-cloud mapping at up to 200+ meters range
- Cameras: Dozens of high-resolution sensors covering visible and infrared spectrums for object classification, traffic signal recognition, and lane detection
- Radar: Short- and long-range radar arrays for velocity detection and redundancy in low-visibility conditions
- Ultrasonic sensors: Close-proximity detection for low-speed maneuvers and pedestrian proximity
- HD Maps + Geofencing: Pre-mapped operational domains with centimeter-level precision; vehicles will not operate outside approved geofence boundaries
- Cloud-connected fleet learning: Real-time data aggregation across the fleet to update perception models and edge-case handling
The Cruise Origin — a purpose-built, steering-wheel-free shuttle designed in partnership with Honda — was intended to be the next-generation robotaxi. It features a spacious, symmetrical interior with no driver controls, sliding doors for curb-side boarding, and an estimated 100+ kWh battery pack targeting 200+ miles of range per charge. However, in mid-2024, GM announced it was pausing Origin production indefinitely, citing regulatory complexity around vehicles without manual driving controls. The company pivoted to using modified Chevrolet Equinox EV platforms for its next testing phase.
The sensor-fusion approach (LiDAR + camera + radar) is widely considered more robust than camera-only for Level 4 operations, particularly in edge cases involving unusual lighting, occluded objects, or sensor degradation. However, it comes at significantly higher per-vehicle hardware cost — estimated at $100,000+ per vehicle for the Cruise Bolt-based platform.
Safety Record: What the Data Actually Shows
Key safety data points:
- Cruise reported over 2.5 million driverless miles completed in San Francisco alone between 2022 and October 2023 (NHTSA filings and Cruise transparency reports)
- California DMV disengagement reports showed Cruise vehicles required approximately 1 intervention per 5,000–10,000 miles during supervised testing phases — comparable to early Waymo figures but significantly higher than Waymo's mature fleet
- The October 2023 incident involved a Cruise vehicle that was struck by a human-driven car; the pedestrian, thrown into the Cruise vehicle's path, was then run over and dragged approximately 20 feet as the vehicle attempted to pull over. Cruise's system detected the collision but did not adequately identify the pedestrian beneath the vehicle
- Following the incident, the California DMV determined Cruise had not adequately represented the safety implications of the event to regulators, leading to permit suspension
- NHTSA opened a formal investigation into Cruise's autonomous driving system under its Standing General Order reporting requirements
For context on comparative safety, human drivers in the U.S. experience approximately 1.35 fatalities per 100 million vehicle miles traveled according to NHTSA data. Robotaxi operators have driven tens of millions of miles collectively without a passenger fatality, but the sample size remains orders of magnitude smaller than the human-driving baseline, making statistical equivalence claims premature.
The Cruise incident highlighted a specific class of edge case — post-collision secondary contact — that autonomous systems across the industry have struggled with. When an external actor (another vehicle, a falling object) introduces a pedestrian into the AV's path at close range, reaction time may be physically insufficient regardless of sensor capability. The distinction is between the system's perception (did it see the pedestrian?) and its prediction/response (did it understand the pedestrian was now under the vehicle and respond appropriately?).
Pricing: How Robotaxis Compare to Uber and Lyft
When Cruise's commercial service was operational in San Francisco (2022–2023), pricing was structured to compete directly with UberX and Lyft standard rides:
| Service | Base Fare | Per-Mile Rate | Avg. 5-Mile Trip | Surge Pricing |
|---|---|---|---|---|
| Cruise (driverless) | $0.00 | ~$1.00–$1.25/mi | ~$8–$12 | No surge |
| Waymo One | $0.00–$2.00 | ~$1.15–$1.50/mi | ~$10–$15 | Limited surge |
| UberX (SF) | $2.50 | ~$1.50–$2.00/mi | ~$14–$22 | Frequent surge |
| Lyft Standard (SF) | $2.00 | ~$1.40–$1.90/mi | ~$13–$20 | Frequent surge |
Cruise aggressively subsidized rides during its commercial phase, offering many trips at $5 or less regardless of distance to build ridership and gather data. This pricing was not sustainable long-term — GM reported losses exceeding $2 billion annually on Cruise operations. The long-term economic model for robotaxis depends on removing the driver cost (typically 60–75% of a rideshare fare) while amortizing the high vehicle hardware cost over millions of revenue miles.
If and when GM resumes commercial driverless service, expect initial pricing to remain competitive with or slightly below UberX, with the potential for significant undercutting during market-capture phases. The absence of surge pricing is a structural advantage — robotaxi fleets manage supply through vehicle repositioning rather than price manipulation.
Regulatory Landscape: Who Approves Robotaxi Operations
Autonomous vehicle deployment in the U.S. involves a patchwork of federal, state, and local approvals:
- Federal (NHTSA): Sets vehicle safety standards (FMVSS). Currently, FMVSS requires manual driving controls (steering wheel, pedals, mirrors) for most vehicle classes — a significant barrier for purpose-built robotaxis like the Cruise Origin. NHTSA also mandates incident reporting through Standing General Orders for all ADS-equipped vehicles.
- State DMVs: Issue deployment and testing permits. California's DMV has the most detailed framework, requiring manufacturers to report disengagements, collisions, and operational metrics. Arizona takes a more permissive approach with minimal state-level oversight.
- State Public Utilities Commissions (PUCs): In California, the CPUC must approve commercial passenger service (charging fares). Cruise held both a DMV deployment permit and a CPUC commercial permit before its suspension.
- Local governments: Cities can impose additional restrictions on AV operations, including hours of operation, geofence boundaries, and fleet-size caps. San Francisco's Board of Supervisors voted to limit robotaxi operations in mid-2023, prior to the October incident.
Following the Cruise suspension, several states have proposed or enacted stricter AV oversight requirements. The regulatory trend in 2025–2026 is toward more transparency, more reporting, and slower expansion approvals — a direct response to both the Cruise incident and growing public concern about AV safety validation.
Current Limitations: What Robotaxis Cannot Do
Even at peak operational capability, Level 4 robotaxis like Cruise face significant constraints:
- Geofencing: Vehicles operate only within pre-mapped areas. Requesting a ride outside the geofence boundary is not possible — the app will not allow booking.
- Weather: Heavy rain, fog, snow, and ice degrade LiDAR and camera performance. Cruise limited or suspended service during significant weather events. Operations in cities with harsh winters (Chicago, Boston) remain impractical with current sensor suites.
- Construction zones: Unmapped lane closures, temporary barriers, and flagger-directed traffic present challenges that frequently require remote assistance or vehicle pull-overs.
- Emergency vehicles: While systems detect sirens and emergency lighting, complex multi-vehicle emergency scenes with unpredictable human behavior remain a known limitation.
- Pickup/dropoff precision: Vehicles pull to the nearest safe curb, which may not be the exact building entrance requested. Riders sometimes need to walk a short distance, particularly on one-way streets or divided roads.
- No highway operation: Cruise's Level 4 service was limited to surface streets at speeds typically under 45 mph. Highway merging and high-speed lane changes were not part of the operational design domain.
- Accessibility: The Bolt-based Cruise vehicle required riders to enter a standard car door. Wheelchair-accessible robotaxi options remain extremely limited industry-wide.
GM's Broader Autonomous Strategy: Super Cruise and Ultra Cruise
While the Cruise robotaxi program resets, GM continues to advance its consumer-facing driver-assistance systems — and it's important not to conflate these with the robotaxi effort:
- Super Cruise (Level 2): Available on select GM vehicles (Cadillac CT5, Escalade; Chevrolet Tahoe, Silverado; GMC Sierra). Enables hands-free driving on 750,000+ miles of pre-mapped compatible roads in the U.S. and Canada. Requires driver attention monitoring via infrared camera. This is not self-driving — the driver must remain alert and ready to take over.
- Ultra Cruise (Level 2): Announced for Cadillac Celestiq and future premium GM models. Extends hands-free capability to virtually all roads (not just highways) using a more advanced sensor suite including LiDAR. Still requires driver supervision. Expected availability in 2026–2027 model-year vehicles.
GM's strategy appears to be pursuing two parallel paths: consumer Level 2+ systems (which generate revenue today and build real-world driving data) and a longer-term Level 4 robotaxi play through Cruise (which remains in a supervised testing and regulatory rebuilding phase).
Frequently Asked Questions
Is GM's self-driving car truly self-driving or just driver-assist?
It depends on which product. The Cruise robotaxi was designed as a true Level 4 self-driving vehicle — no steering wheel, no pedals, no human driver. However, that commercial service is currently suspended. GM's consumer vehicles with Super Cruise and Ultra Cruise are Level 2 driver-assist systems that require the human driver to remain attentive and in control at all times. These are fundamentally different technologies.
Where can I ride in a GM robotaxi today?
You cannot currently ride in a driverless GM/Cruise robotaxi anywhere. Commercial service was suspended in late 2023 and has not been reinstated. GM is conducting supervised testing with safety drivers in select cities. If you want a driverless robotaxi experience today, Waymo One operates commercially in Phoenix, San Francisco, and Los Angeles.
How much did a Cruise robotaxi ride cost compared to Uber?
During its 2023 commercial service in San Francisco, Cruise rides typically cost $8–$12 for a 5-mile trip — generally 30–50% cheaper than comparable UberX or Lyft rides in the same area, and without surge pricing. However, this pricing was heavily subsidized and not representative of a sustainable business model.
What caused Cruise to suspend operations?
In October 2023, a Cruise vehicle in San Francisco struck a pedestrian who had been thrown into its path by a human-driven car. The Cruise vehicle then dragged the pedestrian approximately 20 feet while attempting to pull to the curb. The California DMV suspended Cruise's permit, citing concerns about the company's response to the incident and its representations to regulators. GM voluntarily paused all driverless operations nationwide.
When will GM's robotaxi service return?
GM has not announced a specific date for the return of commercial driverless service. The company has been conducting phased, supervised testing with safety drivers to rebuild operational data and regulatory confidence. Any return to commercial driverless operations will require re-approval from state DMVs and public utilities commissions — a process with no guaranteed timeline.
How does Cruise's technology compare to Waymo's?
Both use sensor-fusion approaches (LiDAR + cameras + radar), but Waymo has accumulated significantly more driverless miles (over 20 million as of 2024) and maintained continuous commercial operations throughout the period when Cruise was suspended. Waymo's fifth-generation hardware is also considered more cost-effective. Cruise's technology demonstrated competent urban driving during its operational period but had a higher intervention rate and faced a more serious safety incident.
