Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/how-tesla-and-waymo%E2%80%99s-radically-different-robotaxi-approaches-will-shape-the-industry.23493/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

How Tesla and Waymo’s radically different robotaxi approaches will shape the industry

Daniel Nenni

Admin
Staff member
1756393925750.png


Tesla and Waymo are pursuing divergent paths in the race to dominate the autonomous ride-hailing market, projected to be a multitrillion-dollar industry within 15 years. A Reuters article from August 28, 2025, details their contrasting approaches, highlighting Tesla's aggressive expansion driven by CEO Elon Musk and Waymo's cautious, safety-first strategy.

Tesla launched a trial robotaxi service in Austin, Texas, in June 2025, initially for select users, with Musk claiming it could reach half the U.S. population by year's end. This ambitious timeline relies on Tesla's vision-based AI, using cameras and neural networks to mimic human driving decisions, bypassing extensive mapping and sensor arrays. Musk argues this approach enables rapid scaling, criticizing Waymo’s method as “fragile” and limited. Tesla’s technology, still in testing, requires less city-specific prep, with Musk stating, “Once we can make it work in a few cities, we can make it work anywhere.” The company is seeking approvals in Arizona, Nevada, and Florida, but faces pressure as its core EV sales decline globally.

Waymo, the U.S. leader in autonomous ride-hailing, operates in five cities—San Francisco, Los Angeles, Phoenix, Austin, and Atlanta—covering about 3% of the population. Its deliberate expansion involves detailed mapping, virtual simulations, and phased testing with safety drivers, taking years per city. For example, Phoenix’s public driverless service began in 2020, over three years after initial tests, with airport service added in August 2024. Waymo’s step-by-step AI integrates high-definition maps and sensors like lidar, prioritizing safety to navigate unique city challenges, such as San Francisco’s steep inclines. Waymo’s product manager, Aman Nalavade, emphasized, “There is a lot of risk in doing this incorrectly.”

Safety concerns highlight the trade-offs. Waymo’s meticulous approach has faced issues, such as vehicles in Austin ignoring police signals or entering floodwaters, prompting citations and safety concerns. Austin police noted Waymo’s vehicles often freeze, blocking traffic, though the company engages with officials to improve. Tesla’s limited Austin rollout has had fewer reported incidents, but a Reuters observer noted a robotaxi speeding near a school for the deaf, raising concerns. Musk insists on caution, stating, “We don’t want to take any chances,” but Tesla’s rapid timeline risks safety oversights.

Regulatory and community challenges further complicate expansion. Waymo’s proactive engagement, like meeting Austin officials and the Texas School for the Deaf over a year before launch, contrasts with Tesla’s minimal outreach, with school officials learning of its launch via news reports. In Washington, D.C., Waymo targets a 2026 launch but faces delays due to pending regulations, with no clear timeline from the city council. Tesla has not contacted D.C.’s transportation department, potentially hindering its plans. Waymo’s lobbying efforts, including hiring firms and circulating petitions, aim to build support, while Tesla’s regulatory hurdles in California—where it lacks permits for paid autonomous rides—underscore its cautious approach despite Musk’s bold claims.

Analysts like Forrester’s Paul Miller view Waymo’s method as safer short-term, while Tesla’s camera-based AI offers cheaper global scaling potential but carries higher risks. Bank of America estimates Waymo’s $1.2-$1.5 billion losses last year, yet foresee profitability as costs drop. Morningstar predicts Tesla’s full robotaxi launch by 2028, overtaking Waymo by decade’s end due to slower initial rollouts. Both companies face a patchwork of U.S. regulations, but Tesla’s aggressive timeline and Waymo’s established operations set the stage for a defining battle in autonomous mobility.

 
This is a very interesting race, though I don't think the approaches are radically different.

On the 'data' side -- Tesla has way more real world miles with self-driving software than anyone including Waymo -- thanks to the many millions of miles of FSD versions driving it's customer cars. BTW here's Tesla pre-mapping Austin just like Waymo did earlier:
On the hardware side, Jim Keller (who designed hardware gen 3 and gen 4) argued with Musk that they needed 5-10X the hardware to make full autonomy a reality. AI5 - not due until Q4 2026 (it was delayed a year) achieves that level of performance, though Tesla is really trying to make it work on the existing HW4/AI4 platform. FSD v14 coming out in September or October will greatly increase the available AI parameter count that can run on HW4 ("up to 10X"), but I'm guessing it's trading off elsewhere to do it.

Waymos approach is VERY expensive compared to Tesla's -- requiring large computers in the trunk (supposedly drawing a few kW - reducing range), but both Tesla and Waymo are premapping streets in all of the areas they're offering services right now.

Tesla is taking better advantage of custom design hardware and TSMC to provide a solution with good cost/power/performance characteristics.

Is Lidar needed? Hard to tell; Lidar does have issues in snow, dust, and fog environments (source: https://www.cts.umn.edu/news/2023/april/lidar) -- while vision clearly works OK in clear conditions, so Lidar might not have an advantage in adverse conditions.

Cost - This is no contest between Waymo and Tesla, at all. Tesla can make it's own Robotaxis for $20-30K each in large volume, that's before you get to the more expensive hardware Waymos need that Teslas currently don't.

Safety is important - but what metrics are the right ones to determine if a mode of unsupervised driving is safe enough overall? Is 2X safer than humans OK? Or does it really need to be 10X so when there is a death caused by an issue, the statistics show than 9 lives were saved in the meantime?
 
Back
Top