Elon Musk believes self-driving cars will handle 90% of all driving within 10 years, but independent research and recent robotaxi failures suggest his timeline is overly optimistic. Speaking at the Samson International Smart Mobility Summit in Tel Aviv this week, the Tesla CEO declared that autonomous vehicles would become so dominant that human driving would become "quite a niche thing" by 2036. However, the gap between Musk's predictions and real-world progress reveals why the autonomous vehicle revolution is moving slower than the billionaire suggests.

What's Actually Holding Back Autonomous Vehicles?

The autonomous driving industry has made genuine progress since artificial intelligence accelerated in 2022, yet massive hurdles remain before self-driving cars can operate reliably on public roads. Recent recalls from both Tesla and Waymo illustrate the persistent technical challenges. Tesla recalled over 200,000 vehicles in the United States due to rearview camera image issues that could increase crash risk, according to the National Highway Traffic Safety Administration (NHTSA). Meanwhile, Waymo recalled approximately 3,800 robotaxis after discovering that vehicles could enter flooded roads at high speeds, a critical safety flaw.

Beyond hardware and software glitches, autonomous systems struggle with what experts call "long tail scenarios," which are unexpected situations that AI systems have never encountered before. Nvidia's vice president of the automotive team Ali Kani explained this challenge to Euronews Next in January, noting that within Musk's predicted five to ten year timeframe, handling these rare but dangerous edge cases remains one of the main obstacles for autonomous vehicles.

A concrete example occurred last year when Waymo's robotaxi service was suspended for hours in San Francisco after vehicles struggled to read malfunctioning stoplights during a power outage. Passengers found themselves stuck at darkened traffic lights with no way to proceed, demonstrating how fragile current autonomous systems can be when conditions deviate from their training data.

When Will Robotaxis Actually Become Mainstream?

According to a 2025 report by the World Economic Forum, a fully autonomous world is unlikely before the mid-2030s at the earliest, and even then adoption will be patchy and concentrated in specific markets. The report found that full autonomy in personal vehicles won't be mainstream by 2035, remaining a niche feature in only 4% of new cars. The closest thing to "mainstream" autonomy will actually be in robotaxis and autonomous trucks, not personal vehicles.

The global robotaxi fleet is projected to grow to somewhere between 700,000 and 3 million vehicles by 2035, but this expansion will be concentrated in just 40 to 80 cities worldwide. This geographic limitation underscores how far autonomous technology still needs to travel before it can handle the diversity of driving conditions across different regions.

How Regulatory Approval Varies Across Regions

The path to autonomous vehicles differs dramatically depending on geography and regulatory approach. Europe currently allows Level 2 systems universally, which means drivers remain responsible for monitoring the road while the vehicle can steer, brake, and accelerate. Europe has already approved Level 3 for controlled conditions, though implementation faces challenges related to regulation and real-world performance.

Some states in the US and China are racing ahead with Level 4 robotaxis, which operate completely autonomously under certain conditions without requiring human intervention. China is expected to adopt higher automation levels fastest, driven by strong consumer appetite and domestic manufacturers like XPeng. However, Level 5 autonomy, which would mean completely driverless operation under all conditions, is explicitly "not currently in sight," according to the International Energy Agency.

"Five years from now and certainly 10 years from now, probably 90% of all distance driven will be driven by the AI in a self-driving car," Elon Musk said at the Samson International Smart Mobility Summit in Tel Aviv.

Elon Musk, CEO at Tesla

Steps to Understanding the Autonomous Vehicle Timeline

• Current Reality: Level 2 and Level 3 autonomous systems are already operational in limited markets, with Level 2 systems requiring driver monitoring and Level 3 approved for controlled highway conditions in Europe.
• Near-Term Growth: Robotaxis operating at Level 4 autonomy will expand in specific cities over the next five years, but full autonomy in personal vehicles remains years away from mainstream adoption.
• Long-Term Outlook: By 2035, autonomous vehicles will likely represent only 4% of new personal cars, with robotaxis and autonomous trucks becoming the primary use cases for self-driving technology.
• Geographic Variation: China is expected to lead in autonomous vehicle adoption, while Europe and the US will move more cautiously due to stricter regulatory frameworks and liability concerns.

While Musk's prediction may be overly optimistic, the autonomous vehicle industry is still advancing. The most realistic scenario involves a gradual expansion of robotaxis alongside human drivers for the next decade, rather than the wholesale replacement of human driving that Musk envisions. Partially autonomous driving is already widespread on roads today, and that incremental progress will likely continue, but the jump to 90% autonomous driving within 10 years remains firmly in the realm of speculation rather than engineering certainty.