Autonomous vehicles have been “a few years away” for a decade. The predictions have consistently been wrong. But dismissing the technology entirely would also be wrong - real progress has happened, even if slower than promised.

Here’s my honest assessment of where autonomous vehicle technology actually stands.

What’s Actually Deployed

Real autonomous vehicles are operating in real environments:

Robotaxis in limited areas. Waymo in San Francisco and Phoenix. Cruise has restarted after setbacks. Others in various cities. These are real services with real passengers - limited geography, but genuine deployment.

Autonomous trucking. Aurora, Kodiak, and others running autonomous freight, initially with safety operators, increasingly without on some routes. Highway driving is more tractable than urban environments.

Closed environments. Mining, ports, warehouses, farms. Controlled environments with predictable conditions and limited variables. This is where autonomous vehicles are most practical today.

ADAS progression. Advanced driver assistance keeps expanding. Not full autonomy, but increasingly capable. Tesla’s Autopilot, GM’s Super Cruise, and others handling more situations.

What’s Still Hard

The challenges that remain are substantial:

Edge cases. The long tail of unusual situations - construction zones, unusual pedestrian behavior, novel obstacles, weather variations. Each edge case can require significant engineering effort.

Sensor reliability. Cameras, LiDAR, radar all have limitations. Weather degrades performance. Sensor failure modes must be handled. Redundancy is expensive.

Human prediction. Anticipating what humans will do - pedestrians, cyclists, other drivers - remains difficult. Humans are unpredictable.

Regulation and liability. Legal frameworks for autonomous vehicles are still developing. Liability questions remain unresolved in many jurisdictions.

Public trust. Every autonomous vehicle incident gets coverage. Trust builds slowly and breaks quickly. The public relations challenge is real.

Economics at scale. Current deployments are expensive per mile. Achieving economic viability at scale requires both cost reduction and volume increase.

The L4/L5 Reality

The SAE levels of automation (0-5) have become misleading:

Level 4 (autonomous in defined conditions) is what’s actually deploying. But “defined conditions” varies dramatically - from a few square miles of San Francisco to controlled mining sites.

Level 5 (autonomous everywhere) remains far off. The “everywhere” part is extremely hard. Every new environment requires extensive mapping and testing.

The practical question isn’t “when will we have L5?” but “which L4 operational design domains are viable, and how fast do they expand?”

Industry Landscape

The competitive landscape has consolidated:

Alphabet/Waymo remains the leader in robotaxis. Years of investment and testing provide advantages.

GM/Cruise has had setbacks but continues development.

Tesla pursues vision-only autonomy, betting on scale and AI. Results remain controversial.

Aurora leads in trucking with a methodical approach.

Chinese players (Baidu, Pony.ai, etc.) are strong in their home market with different regulatory dynamics.

Many others have exited. The difficulty and cost have driven consolidation. Being a standalone AV company is hard.

Investment Perspective

From a technology investment perspective:

Long timelines. Returns are years away for most autonomous vehicle investments. Patient capital is required.

Pick spots carefully. Not all AV applications are equally tractable. Trucking is more viable near-term than urban robotaxis. Closed environments more viable than open roads.

Enable rather than build. Companies supplying sensors, simulation, mapping, and infrastructure for autonomous vehicles may have better risk/return than building complete vehicles.

Watch regulation. Regulatory developments significantly impact which markets and applications are viable.

What I’m Watching

Signals that matter for the trajectory:

Expansion of operational design domains. When Waymo or others expand to new cities or conditions, that’s real progress.

Safety operator removal. The transition from safety operators to truly driverless is the real test.

Unit economics progression. Cost per mile trends indicate whether economic viability is approaching.

Regulatory developments. New regulations that enable or restrict deployment matter significantly.

Insurance market development. Insurance products for autonomous vehicles indicate industry confidence in reliability.

Implications for Organizations

For organizations not in the AV industry, what matters?

Fleet planning. If you operate vehicle fleets, autonomous vehicles will eventually impact operations. Timeline is uncertain but direction is clear.

Logistics strategy. Autonomous trucking will affect logistics economics. Monitor developments.

Real estate implications. Autonomous vehicles could affect parking requirements, urban design, and commercial real estate.

Insurance considerations. The insurance industry will be transformed. Both threats and opportunities exist.

Talent implications. Driving jobs will evolve. Workforce planning should consider this long-term.

Working with AI consultants Brisbane can help organizations assess how autonomous vehicle developments might impact their specific operations and plan accordingly.

Honest Assessment

Autonomous vehicles will happen. The technology works in limited conditions and is expanding.

But the timeline to widespread deployment remains long. Not years - more like decades for full urban autonomy.

The immediate opportunity is in specific, tractable applications: trucking on highways, closed environments, expanding robotaxi operational domains.

For most organizations, autonomous vehicles are something to monitor, not act on immediately. AI consultants Sydney can help with scenario planning, but major strategic moves based on autonomous vehicle assumptions are premature for most industries.

The technology is real. The progress is real. But so is the remaining distance to the autonomous future that’s been promised. Calibrate expectations accordingly.