Tesla Tests Full Self-Driving in Las Vegas Tunnels
The Boring Company has been running supervised tests of Tesla’s Full Self-Driving in the Las Vegas Convention Center tunnels for months, using safety drivers who still must periodically take control. Tunnels simplify some autonomy challenges but introduce new ones—colorful lighting, rock walls and underground pick-ups—that keep widespread deployment a ways off.
Boring Company tests Tesla Full Self-Driving in Las Vegas tunnels
What happened: The Boring Company has been running supervised tests of Tesla’s Full Self-Driving (FSD) inside the tunnels that run beneath the Las Vegas Convention Center, according to statements from Steve Hill, CEO of the Las Vegas Convention and Visitors Authority. Tests have been ongoing for months in Tesla vehicles with safety drivers aboard, but Hill says wider, unsupervised use remains a ways off.
Why this matters: Tunnels are a tempting proving ground for autonomy. They remove many surface variables — traffic, pedestrians, weather — so you might expect the software to handle them more easily. Yet engineers are finding new, tunnel-specific edge cases that complicate perception and navigation.
Real-world context: Tesla has rolled out invite-only robotaxi pilots in Austin and small ride-hail tests in San Francisco, and Musk has touted rapid progress. Still, Hill notes safety drivers must "periodically" intervene in the Las Vegas tunnel runs, signaling that supervised autonomy is not yet hands-off even in this controlled environment.
The specific tunnel challenges include:
- Colorful, dynamic lighting that can confuse camera-based perception.
- Irregular rock or concrete textures on walls that create false features for visual sensors.
- Tight station layouts and underground pick-up/drop-off points that demand precise localization and interaction logic.
- Residual need for human oversight and regulatory approval before public deployment.
Technically, these constraints point to where companies must invest: robust sensor fusion that tolerates unusual lighting, mapping and localization suited for subterranean geometry, and fail-safe behaviors for station operations. Even if the driveable corridor is simple, the margins for error inside a tunnel are tight.
Safety and public perception are central. When a car needs periodic human takeover in a confined environment, that affects how transit agencies, insurers and the public will accept the technology. Think of it like early driverless metro lines: technically promising, but politically and operationally sensitive.
What to watch next: expanded tunnel coverage, intervention rate trends, and whether testing shifts from supervised to supervisory-free operation. Regulators will be tracking metrics, and any permanent deployment will require transparent safety evidence.
How organizations should respond: run focused evaluation programs that measure not just mileage but intervention frequency, environmental edge-case coverage, and passenger board-alight logic. Closed environments are great labs — but they must be instrumented and tested to reveal rare failures before scaling.
QuarkyByte’s approach in scenarios like this is analytical and pragmatic: quantify edge-case exposure, simulate sensor and lighting perturbations, and design pilot metrics that tie directly to operational risk. That combination turns tunnel test runs into clear evidence for regulators and operators considering expansion.
Bottom line: The Boring Company’s use of Tesla FSD inside Las Vegas tunnels is a useful stress test for autonomy. Progress is real, but periodic human interventions and unique subterranean challenges mean cautious, data-driven scaling is still the next mile.
Keep Reading
View AllGoogle Vids Adds AI Avatars and Smart Editing Tools
Google expands Vids with AI avatars, filler-word trimming, image-to-video clips and a free consumer tier for basic editing and templates.
Malaysia Unveils First Domestic Edge AI Processor
SkyeChip launches the MARS1000, Malaysia’s first homegrown edge AI processor, a milestone amid rising AI investment and new chip export permit rules.
Tech Billionaires Back $100M AI Super PAC and Branded Phones
A $100M pro-AI political coalition backed by tech investors reshapes campaign finance while branded MVNOs like Trump Mobile follow celebrity playbooks.
AI Tools Built for Agencies That Move Fast.
QuarkyByte can model tunnel-specific edge cases, simulate lighting and sensor noise, and design safety metrics that show whether autonomy meets operational needs. Contact us to design pilot evaluations, quantify intervention rates, and build a data-driven roadmap for safe, scalable tunnel autonomy.