Walter Isaacson’s Elon Musk biography is set to be published on Tuesday, and a new preview of the book illustrates details about Tesla’s development of the upcoming Full Self-Driving (FSD) version 12.
In an additional preview of his Musk biography for CNBC, Isaacson discusses the use of AI in the development of Tesla’s FSD v12, in a shift that took place within the last several months. Isaacson talks about Tesla’s recent development of the upcoming FSD v12, which he and Tesla demonstrate has moved away from a “rules-based” approach.
Notably, FSD v12 is expected to use billions of video frames from real-world driving incidents to train its neural network rather than using thousands of lines of code like previous versions. In a conversation with Musk last December, Tesla Autopilot employee Dhaval Shroff had likened the concept to the popular chatbot ChatGPT, instead for use with driving.
“It’s like ChatGPT, but for cars,” Shroff said. “We process an enormous amount of data on how real human drivers acted in a complex driving situation, and then we train a computer’s neural network to mimic that.”
Surprisingly enough, Tesla only shifted toward this “neural network planner” approach recently. By the beginning of this year, however, the neural network had already analyzed 10 million video clips based on the best-case-scenario drivers the system had access to. Musk instructed employees at the company’s Buffalo, New York facility who were in charge of analyzing the footage to train the AI on things “a five-star Uber driver would do.”
Moving from a rules-based to a network-path-based AI approach allowed FSD to use human driving data to avoid obstacles, even if breaking some rules was necessary. Shroff helped demonstrate the idea to Musk with a demo featuring trash bins, debris, and upturned traffic cones, which the car handled surprisingly well.
“Here’s what happens when we move from rules-based to network-path-based,” Shroff explained. “The car will never get into a collision if you turn this thing on, even in unstructured environments.”
Musk quickly took to the idea, as can be seen in a recent livestream of Tesla’s FSD v12 software in Palo Alto with Autopilot software director Ashok Elluswamy. He has repeatedly spoken about the upcoming software version’s impressive driving results, despite one small moment in the drive where the car almost ran a red light.
In any case, Musk could argue that the red-light moment is a good case for the need for self-driving software to continually learn. Given that it will constantly be trained from the video data generated by camera footage from real-world drivers, it should theoretically make it safer over time, according to Musk.
During development, Musk also reportedly latched onto the fact that it took over a million video clips for the neural network to begin performing well, though he looks forward to what significantly more data will do for FSD.
Still, critics and regulators have expressed concerns about the faults of human drivers training AI-based driving systems, and Tesla has repeatedly been questioned by the National Highway Traffic Safety Administration (NHTSA) about its Autopilot and FSD beta systems.
According to Isaacson, Tesla plans to release FSD v12 as soon as regulators approve it. Meanwhile, an ongoing study by the National Highway Safety Board is looking to determine if self-driving cars should be permitted to imitate human driving actions that blur traffic rules, such as creeping up at stop signs.
Musk said in April that he expects Tesla to reach full autonomy within a year, though he has also been known to share ambitious targets for the software in the past.
You can read Walter Isaacson’s full account of the development of Tesla FSD v12 here, in a CNBC preview of the upcoming Elon Musk biography.
What are your thoughts? Let me know at zach@teslarati.com, find me on X at @zacharyvisconti, or send your tips to us at tips@teslarati.com.
Investor's Corner
Tesla delivers 384,000 vehicles in Q2 2025, deploys 9.6 GWh in energy storage
The quarter’s 9.6 GWh energy storage deployment marks one of Tesla’s highest to date.
Tesla (NASDAQ: TSLA) has released its Q2 2025 vehicle delivery and production report. As per the report, the company delivered over 384,000 vehicles in the second quarter of 2025, while deploying 9.6 GWh in energy storage. Vehicle production also reached 410,244 units for the quarter.
Model 3/Y dominates output, ahead of earnings call
Of the 410,244 vehicles produced during the quarter, 396,835 were Model 3 and Model Y units, while 13,409 were attributed to Tesla’s other models, which includes the Cybertruck and Model S/X variants. Deliveries followed a similar pattern, with 373,728 Model 3/Ys delivered and 10,394 from other models, totaling 384,122.
The quarter’s 9.6 GWh energy storage deployment marks one of Tesla’s highest to date, signaling continued strength in the Megapack and Powerwall segments.
Year-on-year deliveries edge down, but energy shows resilience
Tesla will share its full Q2 2025 earnings results after the market closes on Wednesday, July 23, 2025, with a live earnings call scheduled for 4:30 p.m. CT / 5:30 p.m. ET. The company will publish its quarterly update at ir.tesla.com, followed by a Q&A webcast featuring company leadership. Executives such as CEO Elon Musk are expected to be in attendance.
Tesla investors are expected to inquire about several of the company’s ongoing projects in the upcoming Q2 2025 earnings call. Expected topics include the new Model Y ramp across the United States, China, and Germany, as well as the ramp of FSD in territories outside the US and China. Questions about the company’s Robotaxi business, as well as the long-referenced but yet to be announced affordable models are also expected.
News
Tesla China breaks 8-month slump by selling 71,599 vehicles wholesale in June
Tesla China’s June numbers were released by the China Passenger Car Association (CPCA) on Tuesday.
Tesla China was able to sell 71,599 vehicles wholesale in June 2025, reversing eight consecutive months of year-over-year declines. The figure marks a 0.83% increase from the 71,599 vehicles sold wholesale in June 2024 and a 16.1% jump compared to the 61,662 vehicles sold wholesale in May.
Tesla China’s June numbers were released by the China Passenger Car Association (CPCA) on Tuesday.
Tesla China’s June results in focus
Tesla produces both the Model 3 and Model Y at its Shanghai Gigafactory, which serves as the company’s primary vehicle export hub. Earlier this year, Tesla initiated a changeover for its best-selling vehicle, the Model Y, resulting in a drop in vehicle sales during the first and second quarters.
Tesla’s second-quarter China sales totaled 191,720 units including exports. While these numbers represent a 6.8% year-over-year decline for Tesla China, Q2 did show sequential improvement, rising about 11% from Q1 2025, as noted in a CNEV Post report.
For the first half of the year, Tesla sold 364,474 vehicles wholesale. This represents a 14.6% drop compared to the 426,623 units sold wholesale in the first half of 2024.
China’s competitive local EV market
Tesla’s position in China is notable, especially as the new Model Y is gaining ground in the country’s BEV segment. That being said, Tesla is also facing competition from impressive local brands such as Xiaomi, whose new YU7 electric SUV is larger and more affordable than the Model Y.
The momentum of the YU7 is impressive, as the vehicle was able to secure 200,000 firm orders within three minutes and over 240,000 locked-in orders within 18 hours. Xiaomi’s previous model, the SU7 electric sedan, which is aimed at the Tesla Model 3, also remains popular, with June deliveries surpassing 25,000 units for the ninth straight month.
While China’s EV market is getting more competitive, Tesla’s new Model Y is also ramping its production and deliveries. Needless to say, Tesla China’s results for the remaining two quarters of 2025 will be very interesting.
Elon Musk
Tesla reveals it is using AI to make factories more sustainable: here’s how
Tesla is using AI in its Gigafactory Nevada factory to improve HVAC efficiency.
Tesla has revealed in its Extended Impact Report for 2024 that it is using Artificial Intelligence (AI) to enable its factories to be more sustainable. One example it used was its achievement of managing “the majority of the HVAC infrastructure at Gigafactory Nevada is now AI-controlled” last year.
In a commitment to becoming more efficient and making its production as eco-friendly as possible, Tesla has been working for years to find solutions to reduce energy consumption in its factories.
For example, in 2023, Tesla implemented optimization controls in the plastics and paint shops located at Gigafactory Texas, which increased the efficiency of natural gas consumption. Tesla plans to phase out natural gas use across its factories eventually, but for now, it prioritizes work to reduce emissions from that energy source specifically.
It also uses Hygrometric Control Logic for Air Handling Units at Giafactory Berlin, resulting in 17,000 MWh in energy savings each year. At Gigafactory Nevada, Tesla saves 9.5 GWh of energy through the use of N-Methylpyrrolidone refineries when extracting critical raw material.
Perhaps the most interesting way Tesla is conserving energy is through the use of AI at Gigafactory Nevada, as it describes its use of AI to reduce energy demand:
“In 2023, AI Control for HVAC was expanded from Nevada and Texas to now include our Berlin-Brandenburg and Fremont factories. AI Control policy enables HVAC systems within each factory to work together to process sensor data, model factory dynamics, and apply control actions that safely minimize the energy required to support production. In 2024, this system achieved two milestones: the majority of HVAC infrastructure at Gigafactory Nevada is now AI-controlled, reducing fan and thermal energy demand; and the AI algorithm was extended to manage entire chiller plants, creating a closed-loop control system that optimizes both chilled water consumption and the energy required for its generation, all while maintaining factory conditions.”
Tesla utilizes AI Control “primarily on systems that heat or cool critical factory production spaces and equipment.” AI Control communicates with the preexisting standard control logic of each system, and any issues can be resolved by quickly reverting back to standard control. There were none in 2024.
Tesla says that it is utilizing AI to drive impact at its factories, and it has proven to be a valuable tool in reducing energy consumption at one of its facilities.
Tesla delivers 384,000 vehicles in Q2 2025, deploys 9.6 GWh in energy storage
Tesla China breaks 8-month slump by selling 71,599 vehicles wholesale in June
Tesla reveals it is using AI to make factories more sustainable: here’s how
-
Elon Musk2 days agoTesla investors will be shocked by Jim Cramer’s latest assessment
-
News7 days agoTesla Robotaxi’s biggest challenge seems to be this one thing
-
News2 weeks agoTesla’s Grok integration will be more realistic with this cool feature
-
Elon Musk2 weeks agoElon Musk slams Bloomberg’s shocking xAI cash burn claims
-
News2 weeks agoTexas lawmakers urge Tesla to delay Austin robotaxi launch to September
-
News2 weeks agoTesla dominates Cars.com’s Made in America Index with clean sweep
-
Elon Musk1 week agoFirst Look at Tesla’s Robotaxi App: features, design, and more
-
Elon Musk2 weeks agoTesla Robotaxis are becoming a common sight on Austin’s public roads