Every quarter, Waymo updates its Safety Impact page with the latest Rider-Only (RO) miles driven across its operating cities: Los Angeles, San Francisco, Phoenix, and Austin. Here’s the latest snapshot released in March 2026:
Really good analysis. The only thing that caught my atention was the 70% market share of Waymo in PHX in jun 2024. If I get it correct, Waymo had 70% of all ride sharing in PHX? From where are you taking the total miles in every market? Or just an estimate?
I think he's just using the term "market share" incorrectly. He's attempting to show that Phoenix accounts for a smaller percentage of all Waymo miles than it used to.
Fair, it's a little confusing, and I wasn't sure what the best term would be to use. But essentially it's just 'Waymo's market share'. So of all Waymo's RO miles across all 4 cities, Phoenix had 70% of those miles (ie market share) in June of 2024. By Dec. of 2025, Phoenix was just 40% of Waymo's total miles, meaning the other markets are now doing more miles and catching up to Phoenix.
Do you have a sense of why Waymo doesn't report the Atlanta number even though it reported the Austin number even at a comparatively small mileage? (<1M)
Good question, public rides began in Austin in March 2025, and Atlanta in June of 2025, so it's only one quarter behind, yet we have 6 quarters of Austin data.
According to Waymo, "The comparisons in Atlanta are not shown here due to Waymo’s limited mileage, which means the results are not yet statistically significant."
Atlanta also has ~100 vehicles, while Austin has ~200, so I would imagine worst case, Atlanta is at half or 1/3 of the Austin mileage, so doesn't make sense to not include it imo. Maybe something for someone to dig into further :)
I wish that other companies would distill their data and be more open in guiding the public on the safety of their program. Some companies instead go as far as to mix and match their miles in places with riders, observers and the occasional rider only miles go into one flaming pile. Those statistics are irrelevant at best and downright deceptive at worst.
One of the coolest takeaways from the safety report is how much effort it takes to get to safety relevant (about 10M rider only miles). Only these four cities got there so far. The biggies are beyond 100K miles a day and Austin closer to 30K per day. This what it takes to reveal as a 'serious' market. In Waymo's world, it would be IRRESPONSIBLE to protect the inherent safety in a market until they get to about 10M miles. It is refreshing to treat statistics responsibly. That is what it takes to draw real safety conclusions based on error bars.
It will be interesting to watch how long before new cities emerge. I am guessing that Miami will be the next big market and maybe DFW and get to 10M miles rather quickly -- maybe as early as 3Q this year so by the EOY safety report. Phoenix, San Francisco & Los Angeles are now past 100K miles per day while Austin trails at 30K/day. I think Miami will get there surprisingly quickly -- airport service is a tell b/c that's a big part of taxi service.
I am glad Zoox is at least providing guidance in RO miles and now is past 2M miles. If they meet their goals for this year they have outlined they will emerge close to 10M between 4Q26 and 2Q27. Still early days but at least they are providing useful guidance. It is intentionally difficult to guess where Tesla is at. I guess they are likely somewhere between 75 miles/day when Elon & Ashok rode in January and maybe approaching 3000 miles/week after 3+ months now. Still only 10a-3p when not raining and MAYBE 3 cars. All of the rest of the 'meaningful miles' are silly and irrelevant. Providing real RO miles is hard. It is a great shortcut to identify the shill. That's 3000 miles a week in a good week I would guess. Irresponsible to even guess how long to get 10M miles of experience in Austin in even a modest ODD. Time will tell.
Yea I think it's commendable that Waymo is releasing this much data, even though I have criticized them for not releasing enough data when it comes to incidents like the SF outage, first responder issues, etc in the past.
Part of me did worry though that by writing this, they may stop providing this level of granular data. Hopefully not.. :)
Really good analysis. The only thing that caught my atention was the 70% market share of Waymo in PHX in jun 2024. If I get it correct, Waymo had 70% of all ride sharing in PHX? From where are you taking the total miles in every market? Or just an estimate?
Anyways, really helpful.
I think he's just using the term "market share" incorrectly. He's attempting to show that Phoenix accounts for a smaller percentage of all Waymo miles than it used to.
Yea fair, I'm not sure what the best term would be but open to ideas/suggestions.
I like market diversity. To showcase how quickly waymo is expanding across cities.
Market share should be used to compare after Zoox and Tesla or Uber+Lyft vs self driving vs taxis.
Market share of just waymo's share of the market is a very weird graph
Fair, it's a little confusing, and I wasn't sure what the best term would be to use. But essentially it's just 'Waymo's market share'. So of all Waymo's RO miles across all 4 cities, Phoenix had 70% of those miles (ie market share) in June of 2024. By Dec. of 2025, Phoenix was just 40% of Waymo's total miles, meaning the other markets are now doing more miles and catching up to Phoenix.
Nice writeup.
Do you have a sense of why Waymo doesn't report the Atlanta number even though it reported the Austin number even at a comparatively small mileage? (<1M)
Good question, public rides began in Austin in March 2025, and Atlanta in June of 2025, so it's only one quarter behind, yet we have 6 quarters of Austin data.
According to Waymo, "The comparisons in Atlanta are not shown here due to Waymo’s limited mileage, which means the results are not yet statistically significant."
Atlanta also has ~100 vehicles, while Austin has ~200, so I would imagine worst case, Atlanta is at half or 1/3 of the Austin mileage, so doesn't make sense to not include it imo. Maybe something for someone to dig into further :)
https://www.thedriverlessdigest.com/p/waymo-stats-2025-funding-growth-coverage
I wish that other companies would distill their data and be more open in guiding the public on the safety of their program. Some companies instead go as far as to mix and match their miles in places with riders, observers and the occasional rider only miles go into one flaming pile. Those statistics are irrelevant at best and downright deceptive at worst.
One of the coolest takeaways from the safety report is how much effort it takes to get to safety relevant (about 10M rider only miles). Only these four cities got there so far. The biggies are beyond 100K miles a day and Austin closer to 30K per day. This what it takes to reveal as a 'serious' market. In Waymo's world, it would be IRRESPONSIBLE to protect the inherent safety in a market until they get to about 10M miles. It is refreshing to treat statistics responsibly. That is what it takes to draw real safety conclusions based on error bars.
It will be interesting to watch how long before new cities emerge. I am guessing that Miami will be the next big market and maybe DFW and get to 10M miles rather quickly -- maybe as early as 3Q this year so by the EOY safety report. Phoenix, San Francisco & Los Angeles are now past 100K miles per day while Austin trails at 30K/day. I think Miami will get there surprisingly quickly -- airport service is a tell b/c that's a big part of taxi service.
I am glad Zoox is at least providing guidance in RO miles and now is past 2M miles. If they meet their goals for this year they have outlined they will emerge close to 10M between 4Q26 and 2Q27. Still early days but at least they are providing useful guidance. It is intentionally difficult to guess where Tesla is at. I guess they are likely somewhere between 75 miles/day when Elon & Ashok rode in January and maybe approaching 3000 miles/week after 3+ months now. Still only 10a-3p when not raining and MAYBE 3 cars. All of the rest of the 'meaningful miles' are silly and irrelevant. Providing real RO miles is hard. It is a great shortcut to identify the shill. That's 3000 miles a week in a good week I would guess. Irresponsible to even guess how long to get 10M miles of experience in Austin in even a modest ODD. Time will tell.
Yea I think it's commendable that Waymo is releasing this much data, even though I have criticized them for not releasing enough data when it comes to incidents like the SF outage, first responder issues, etc in the past.
Part of me did worry though that by writing this, they may stop providing this level of granular data. Hopefully not.. :)