# Why Is Waymo Still the Only Company Publishing Safety Data?

According to Bloomberg's June 28, 2026 feature, Waymo has released quarterly safety reports since 2023 that detail every collision involving its autonomous vehicles, including those caused by human drivers. The company reported 37 total collisions in 2025 across its Phoenix and San Francisco fleets, with zero at-fault incidents. Tesla, by contrast, has not published any comparable dataset. Elon Musk claimed during the 2025 shareholder meeting that Tesla's Full Self-Driving (FSD) software is "safer than a human driver by a factor of 10," but provided no raw data to support the assertion. This asymmetry matters because regulators in California and Arizona now require annual safety reporting from all autonomous vehicle operators — but Tesla's vehicles are still classified as Level 2 driver-assist, exempting it from the same scrutiny. The result is a regulatory loophole that allows Tesla to market autonomy without proving it.

# What Does Baidu's Apollo Go Actually Deliver in China?

Bloomberg reported that Baidu's Apollo Go robotaxi service completed 2.1 million paid rides in 2025, up from 600,000 in 2024, operating across 10 Chinese cities including Beijing, Wuhan, and Guangzhou. However, each city requires separate regulatory approval, and Baidu's expansion has been slower than the company projected in 2024. The service is also restricted to geofenced zones that exclude highways and complex urban intersections. "We are not yet at the point where Apollo Go can operate everywhere," a Baidu spokesperson told Bloomberg. "The regulatory process is deliberate by design." This places Baidu in a middle ground: ahead of Tesla in deployed units, but behind Waymo in safety documentation and operational transparency.

# How Do the Three Competitors Compare on Key Metrics?

# Why Are Glitches Still a Problem After So Much Testing?

Bloomberg documented several high-profile glitches in 2025 and early 2026. In November 2025, a Waymo vehicle in San Francisco stopped mid-block for 12 minutes because it could not determine whether a plastic bag was a pedestrian or debris. In March 2026, a Tesla on FSD in Los Angeles failed to recognize a construction zone and had to be manually overridden. Baidu's Apollo Go vehicles have been reported to struggle with heavy rain and snow, leading to service suspensions during inclement weather. According to Dr. Sarah Chen, a robotics researcher at MIT cited in Bloomberg's piece, "The long tail of edge cases is the fundamental obstacle. Every company solves 99% of driving scenarios, but that remaining 1% contains infinite variations." This explains why no company has yet deployed a truly weather-agnostic, nationwide robotaxi service.

# Who Gains and Who Loses From This Uneven Progress?

Gains: Waymo's parent company Alphabet gains credibility with regulators and insurers, which translates into lower insurance premiums and faster permitting. Baidu gains political capital in Beijing by demonstrating Chinese technological leadership. Consumers in Phoenix and San Francisco gain a reliable, if limited, transportation option.

Loses: Tesla loses investor confidence with each missed deadline. Uber and Lyft lose pricing power as robotaxis undercut human-driven rides in geofenced areas. The general public loses trust in autonomous technology when high-profile glitches make headlines, slowing adoption everywhere.

# What Remains Genuinely Uncertain?

Three unknowns stand out: First, whether any company can achieve Level 5 autonomy (no geofence, no driver) within the next decade. Second, whether the regulatory patchwork in the US and EU will harmonize or fragment further. Third, whether public acceptance will grow as glitches become less frequent, or whether a single fatal incident could set the industry back years. Bloomberg's article does not answer these questions, and neither do the companies themselves.

1. By December 2027, Waymo will have expanded to at least five US cities (adding Los Angeles and Austin) and will maintain its lead in safety reporting.
2. By June 2028, Tesla will not have launched a driverless robotaxi service in any US city, despite Musk claiming it will happen "next year" in 2027.
3. By December 2028, Baidu's Apollo Go will operate in 15 Chinese cities but will still be unprofitable, requiring continued state subsidies.

• 2023: Waymo begins publishing quarterly safety reports.
• 2024: Baidu's Apollo Go completes 600,000 paid rides.
• 2025: Waymo reports 37 collisions, zero at-fault. Baidu completes 2.1 million rides. Tesla FSD beta remains Level 2.
• June 2026: Bloomberg publishes feature on robotaxi progress and glitches.

• The robotaxi market is not a single race — it's three different competitions in three different regulatory environments.
• Safety transparency is the single most important differentiator for long-term trust, and only Waymo is investing in it publicly.
• Tesla's brand alone cannot overcome the absence of verifiable safety data; investors should discount its autonomy timeline by at least three years.
• Baidu's volume lead is real but fragile, dependent on government goodwill and subsidies that could shift with political winds.
• The edge-case problem is not solvable by scaling alone; it requires fundamentally new approaches to perception and decision-making that no company has demonstrated.