Autonomous vehicles have been “almost here” for a decade. Promises of robotaxi fleets have repeatedly missed deadlines. Yet genuine progress continues beneath the hype cycles.

I’ve been tracking AV development since the DARPA challenges. Here’s an honest assessment of where we stand.

Current Deployment Reality

Autonomous vehicles exist in commercial operation today, with constraints:

Waymo: Operating robotaxi services in Phoenix, San Francisco, Los Angeles, and Austin. Genuinely driverless vehicles handling real passengers. The most advanced commercial deployment.

Cruise: Paused operations after a 2023 incident, working to resume with enhanced safety measures.

Zoox: Testing robotaxis in limited areas. Amazon backing provides runway for continued development.

Tesla FSD: Widely deployed driver assistance that requires constant supervision. Not truly autonomous despite the name.

Chinese players: Baidu, Pony.ai, and others operating robotaxis in Chinese cities. Progress harder to assess from outside.

Trucking: Aurora, Kodiak, and others running autonomous trucks on specific highway routes with remote supervision.

What’s Working

Autonomous technology has advanced substantially:

Perception: Modern AV perception systems are remarkably capable. LiDAR, cameras, and radar combined with AI create detailed environment understanding.

Highway driving: Long-haul trucking on predictable highway routes is approaching commercial viability.

Geofenced urban operation: Robotaxis work in carefully mapped, controlled areas with good conditions.

Structured environments: Ports, warehouses, mining operations—AVs succeed in controlled settings.

What Remains Hard

Significant challenges persist:

Edge cases: Rare but critical situations that humans navigate easily. Construction zones, accident scenes, unusual road conditions.

Weather: Rain, snow, fog, and bright sun all degrade AV performance.

Human interaction: Predicting and responding to unpredictable human behavior—pedestrians, cyclists, aggressive drivers.

Mapping dependency: Many AVs require detailed pre-mapped areas. Scaling mapping to everywhere is expensive.

Long-tail safety: Getting from 99.9% safe to 99.999% safe is harder than the previous improvements combined.

The Business Model Question

Even where technology works, economics remain challenging:

Capital intensity: AV development requires billions. Only a few players have necessary resources.

Per-vehicle costs: Current AVs cost $150,000-$300,000 per vehicle. Scale economics eventually help but aren’t there yet.

Utilization requirements: To be economical, robotaxis must operate many more hours than personal vehicles.

Regulatory fragmentation: Different rules in different jurisdictions complicate scaling.

Public trust: Each incident, however rare, damages public perception and invites regulatory response.

Industry Consolidation

The AV industry has contracted from peak hype:

Exits: Uber, Lyft sold their AV units. Argo AI shut down. Apple reportedly scaled back.

Survivors: Waymo, Tesla, Amazon-backed Zoox, a handful of Chinese players.

Pivots: Some AV startups now focus on ADAS features rather than full autonomy.

Consolidation continues: Expect more departures and acquisitions.

Realistic Timeline

Where autonomous vehicles are headed:

Now: Limited robotaxi operations in select cities. Commercial truck pilots on specific routes.

2026-2027: Expanded robotaxi footprint in favorable markets. Initial commercial trucking operations.

2028-2030: Robotaxis in dozens of cities worldwide. Substantial autonomous trucking deployment.

2031-2035: Personal vehicles with meaningful autonomous capability in some conditions.

2035+: Widespread autonomous operation, though likely not universal.

This timeline assumes continued progress without major setbacks.

Implications for Adjacent Industries

AV development affects connected sectors:

Insurance: Liability shifts from driver to manufacturer. New insurance models required.

Automotive OEMs: Traditional manufacturers must adapt or face disruption.

Real estate: If commuting becomes productive time, distance from work matters less.

Urban planning: Parking needs decrease; street design changes.

Employment: Driving jobs at risk, but timeline longer than often predicted.

What to Watch

Signals worth monitoring:

Waymo expansion: If Waymo continues adding cities successfully, full autonomy is viable.

Tesla FSD evolution: Can Tesla close the gap between driver assistance and true autonomy?

Regulatory changes: Governments either enabling or restricting AV deployment.

Incident trends: Safety record improving or plateauing?

Cost curves: Are per-vehicle costs declining at scale?

My Take

Autonomous vehicles are real but overhyped. They work in constrained conditions today and will work in more conditions over time. The path to ubiquitous autonomy is longer than enthusiasts claim but real.

For individuals, AVs remain a distant consideration for personal transportation. For businesses, robotaxi services are becoming usable in select cities. For investors, extreme caution given capital requirements and uncertain timelines.

The technology is impressive. Making it work everywhere, safely, economically—that’s a longer journey.

---

Tracking the slow but steady progress toward autonomous transportation.