Google’s driverless ride-hailing company Waymo has revealed the latest in a string of AI research-related announcements, this time sharing results from a first-of-its-kind study on collisions with vulnerable road users.
Waymo last week announced EMMA, its end-to-end, multimodal research model for autonomous driving, though it isn’t being used commercially at this point. The company also went on to detail its current approach to AI in a separate press release, noting its continued research into real-world models and AI training, a concept that may sound familiar to those who have followed Tesla’s Full Self-Driving (FSD) development.
On Monday, however, Waymo shared results from a study on Vulnerable Road Users (VRUs), in which the company reconstructed hundreds of collisions involving VRUs like pedestrians, cyclists, and motorcyclists. Waymo says that the research, as conducted in a partnership with Nexar, is the largest dataset of its kind in the U.S., providing key insights into real-world crash scenarios.
The companies analyzed dash camera footage of 335 collisions involving VRUs across six U.S. cities, leveraging over 500 million miles of driving data from Nexar for the research. The partnership also worked with Waymo research partner VUFO, which contributed to the below models on collision injury risk.
Credit: Waymo Credit: Waymo

According to the National Highway Traffic Safety Administration (NHTSA), there were an estimated 7,522 pedestrians and 1,105 cyclists who lost their lives in traffic crashes in 2022, though Waymo says VRU collision data remains scarce. The company explains this by noting that several VRU-related incidents don’t get reported to authorities or insurance firms, while the U.S. and other major accident databases lack data on these particular incidents.
“We’re excited to partner with Waymo on this cutting-edge research,” said Henrik Liers, Managing Director of Waymo research partner VUFO. “Our common work addresses critical limitations in existing models and provides a more precise, interdisciplinary approach to assessing injury risk for vulnerable road users. This is a significant contribution towards improving road safety.”
Tesla’s Full Self-Driving and end-to-end learning models
The recent developments from Waymo come as General Motors’s (GM’s) Cruise, Amazon’s Zoox, and Tesla’s Supervised Full Self-Driving (FSD) aim to offer similar robotaxi services. While Tesla doesn’t currently operate a paid ride-hailing service as Waymo does, it unveiled the two-seat, steering wheel-free Cybercab robotaxi last month.
Interestingly, Tesla’s FSD system has been touted by some as a more scalable solution, in no small part due to its end-to-end system that trains on millions of clips of real-time driving footage. Another factor is that FSD is available to any Tesla owner who purchases the software, meaning that its neural network stands to have a much wider potential research base than systems like Waymo—at least until services scale up substantially.
Tesla’s Cybercab isn’t expected to go into production until 2026, though FSD Supervised will certainly collect a substantial amount of data in the meantime. As of Tesla’s Q3 earnings call, the company has over 2 billion cumulative miles of data from FSD Supervised users, after it surpassed a milestone of 1.3 billion miles in April.
Tesla FSD V13 to implement features required for unsupervised driving: exec
What are your thoughts? Let me know at zach@teslarati.com, find me on X at @zacharyvisconti, or send us tips at tips@teslarati.com.
Need accessories for your Tesla? Check out the Teslarati Marketplace:
Investor's Corner
Bank of America raises Tesla PT to $471, citing Robotaxi and Optimus potential
The firm also kept a Neutral rating on the electric vehicle maker, citing strong progress in autonomy and robotics.
Bank of America has raised its Tesla (NASDAQ:TSLA) price target by 38% to $471, up from $341 per share.
The firm also kept a Neutral rating on the electric vehicle maker, citing strong progress in autonomy and robotics.
Robotaxi and Optimus momentum
Bank of America analyst Federico Merendi noted that the firm’s price target increase reflects Tesla’s growing potential in its Robotaxi and Optimus programs, among other factors. BofA’s updated valuation is based on a sum-of-the-parts (SOTP) model extending through 2040, which shows the Robotaxi platform accounting for 45% of total value. The model also shows Tesla’s humanoid robot Optimus contributing 19%, and Full Self-Driving (FSD) and the Energy segment adding 17% and 6% respectively.
“Overall, we find that TSLA’s core automotive business represents around 12% of the total value while robotaxi is 45%, FSD is 17%, Energy Generation & Storage is around 6% and Optimus is 19%,” the Bank of America analyst noted.
Still a Neutral rating
Despite recognizing long-term potential in AI-driven verticals, Merendi’s team maintained a Neutral rating, suggesting that much of the optimism is already priced into Tesla’s valuation.
“Our PO revision is driven by a lower cost of equity capital, better Robotaxi progress, and a higher valuation for Optimus to account for the potential entrance into international markets,” the analyst stated.
Interestingly enough, Tesla’s core automotive business, which contributes the lion’s share of the company’s operations today, represents just 12% of total value in BofA’s model.
News
Tesla Model Y on FSD saves couple after encountering King of edge cases
Experts have noted that if confirmed, this could be the world’s first recorded meteorite collision involving a Tesla.
A South Australian Tesla driver is thanking his Model Y’s Full Self-Driving system after a mysterious object, possibly a meteorite, slammed into his car’s windshield while it had FSD engaged.
The impact sent hot glass fragments flying through the cabin as the vehicle continued driving without human input through the darkness. Experts have noted that if confirmed, this could be the world’s first recorded meteorite collision involving a Tesla.
The Tesla owner was enjoying a quiet drive home when they hit the king of edge cases
Veterinarian Dr. Andrew Melville-Smith and his wife were traveling north on Augusta Highway on the night of October 19 when a sudden blast struck their newly delivered Model Y. At the time, it was clear, pitch black night, and Dr. Melville-Smith and his wife were just listening to a podcast while FSD was operating the vehicle, as noted in a Yahoo News report.
Suddenly, something hit the Tesla’s windshield. “Then there was, (what) I can only describe as a very, very violent explosion. The whole inside of the car was literally blasted with glass fragments. It was full of white smoke, and it smelled like the car was on fire,” Dr. Melville-Smith stated.
The vet stated that he and his wife were stunned for a good 10 seconds, and for a bit, he and his wife thought they had crashed. Images of the crash’s aftermath showed a massive crack on the Tesla’s windshield, which Dr. Melville-Smith noted was hot to the touch. The object that hit the Tesla’s windshield was so hot that part of the vehicle’s windshield was partly melted.
FSD kept driving despite the possible meteorite strike
Thankfully, the Tesla was operating with its FSD (Supervised) system engaged at the time. Thus, despite the vehicle being hit by what might be a literal meteorite, and despite Dr. Melville-Smith and his wife being stunned because of the impact, their Tesla just kept driving steadily.
“I thought we’d had an accident, but then I looked at the screen and went ‘Oh, we’re still driving. We were moving around, so it obviously thought we were paying attention, and it was happy to keep driving to Port Augusta. After we pulled over, we saw the big crater in the windscreen,” the vet said.
The South Australian Museum is now examining the case, with geologist Dr. Kieran Meaney stating the scorched glass and heat damage do suggest a potential meteorite strike. “The little detail that’s really selling it for me at the moment is that whatever it was that hit the windscreen seems to have been very hot,” Meaney explained. However, the object has yet to be recovered.
Check out a video of the remarkable edge case below.
News
“Foundation:” Elon Musk wants to send a record of Grokipedia to space
The idea sounds outlandish, though it is also something that is in character for the CEO.
Elon Musk has announced one of his latest initiatives, and it is every bit as sci-fi as it is surprisingly grounded. In a post on X, Musk stated that Grokipedia, xAI’s new open-source encyclopedia, will be etched into stable oxide and launched into space.
The idea sounds outlandish, though it is also something that is in character for the CEO.
Preserving human knowledge among the stars
Musk posted his plan following the launch of Grokipedia’s V0.1’s iteration. The CEO congratulated the xAI team for the online encyclopedia’s launch, though he also stated that the goal for Grokipedia is to create an open-source collection of knowledge. This would then be distributed to the cosmos.
“Nice work by the xAI team on Grokipedia! The goal here is to create an open source, comprehensive collection of all knowledge. Then place copies of that etched in a stable oxide in orbit, the Moon, and Mars to preserve it for the future. Foundation,” Musk wrote in his post.
While seemingly outlandish, this is not the first time that a record of human knowledge of sorts was sent out to space. In 1977, the Voyager Golden Record was launched aboard NASA’s Voyager 1 and 2 spacecraft. The record contains sounds and images that ere aggregated to portray the diversity of the Earth’s culture. Of course, Musk’s plan with Grokipedia, is infinitely more ambitious.
Grokipedia and AI neutrality
Musk launched Grokipedia as an AI-driven alternative to Wikipedia, designed to eliminate the human biases that could affect conventional online knowledge platforms. The system is powered by xAI’s Grok, which scrape and summarize information from across the internet, offering balanced and nuanced coverage of topics ranging from science and technology to culture and politics.
Unlike Wikipedia’s human-edited format, Grokipedia would be able to evolve through machine learning, reading vastly more material than any editorial team could. Early testers, including Wikipedia co-founder Larry Sanger, praised its initial version as “very OK” and potentially more neutral than Wikipedia. Musk agreed, stating that even in its V0.1 form, Grokipedia is “already better than Wikipedia.”
-
Elon Musk2 weeks agoSpaceX posts Starship booster feat that’s so nutty, it doesn’t even look real
-
Elon Musk2 weeks agoTesla Full Self-Driving gets an offer to be insured for ‘almost free’
-
News2 weeks agoElon Musk confirms Tesla FSD V14.2 will see widespread rollout
-
News2 weeks agoTesla is adding an interesting feature to its centerscreen in a coming update
-
News2 weeks agoTesla launches new interior option for Model Y
-
News2 weeks agoTesla widens rollout of new Full Self-Driving suite to more owners
-
Elon Musk2 weeks agoTesla CEO Elon Musk’s $1 trillion pay package hits first adversity from proxy firm
-
News1 week agoTesla might be doing away with a long-included feature with its vehicles

