Tesla (NASDAQ:TSLA) reported record third-quarter revenue of 28.1 billion USD for Q3 2025, surpassing Wall Street forecasts but falling short on profit expectations .
Tesla Q3 2025 earnings and webcast recap: what Elon Musk and team revealed:
8:57 – Tesla Q3 2025 Q&A webcast begins
The webcast opens with a welcome from Travis Axelrod, Tesla’s head of investor relations, alongside Elon Musk and other executives. They reference the announcement of Tesla’s third quarter 2025 results and the accompanying update deck. The discussion will cover the business outlook and include forward-looking statements based on current predictions and expectations. They caution that actual results may vary due to risks and uncertainties detailed in recent SEC filings. Shareholders are encouraged to review the definitive proxy statement for important voting information. During the Q&A session, participants are asked to limit themselves to one question and one follow-up.
10:50 – Elon Musk on Tesla AI and full self-driving
Elon Musk begins with opening remarks emphasizing Tesla’s leadership in real-world AI, particularly in the intelligence embedded in their cars. He highlights the transformative potential of Tesla’s full self-driving technology and robotaxi services, predicting a rapid and significant impact on transportation. Elon notes that millions of Tesla vehicles can be upgraded to full self-driving via software updates, and with recent clarity on achieving unsupervised full self-driving, Tesla plans to expand production aggressively. Additionally, Tesla is making significant advances in energy storage solutions, notably with Powerwall and Megapack, to enhance energy generation from the grid.
13:35 – Tesla’s battery storage impact and energy strategy
The speaker discusses the energy capacity in the US, noting that although there is roughly a terawatt of continuous power available, average usage is only half a terawatt due to daily fluctuations. By using batteries to store energy, the effective energy output can be doubled without building new power plants, which are difficult and slow to develop due to permitting and industry constraints. Tesla’s battery technology has significant potential to improve yearly energy output for grids. Additionally, the speaker mentions the promising development of Optimus, a challenging project that could become Tesla’s biggest product ever.
14:52 – Optimus robot development and challenges
Elon Musk discusses Tesla’s approach to robotics, emphasizing that while no major US car companies have clear robot programs, Tesla has the necessary AI, engineering skills, and production capabilities to develop Optimus. He highlights the release of version 14 of Tesla’s self-driving software, which is available to all US users who select the advanced software option, and mentions positive reactions from the public.
16:20 – Tesla unveiled the Mega Pack
Tesla unveiled the Mega Pack 3 energy storage system and is developing Mega Pack 4, which will improve deployment by integrating substation components and outputting at 35 kilovolts directly, removing the need for separate substations. This advancement is a key engineering priority for Tesla’s energy solutions.
17:08 – Elon Musk previews Optimus V3
Elon Musk previews Optimus V3, expected to debut in Q1, describing it as so lifelike it will resemble a person in a robot suit. Much of the real-world intelligence developed for Tesla’s cars will transfer to Optimus, providing a strong foundation. Musk concludes by sharing Tesla’s mission of sustainable abundance, envisioning a future where technologies like Optimus and self-driving eliminate poverty and provide universal access to top medical care.
18:43 – Musk envisions Optimus as an incredible surgeon
Musk envisions Optimus as an incredible surgeon accessible to all, contributing to a future of sustainable abundance. He stresses the importance of safety in development and expresses enthusiasm for Tesla’s mission. The segment ends with Vebup acknowledging Q3 as a record-setting quarter for Tesla in deliveries, deployments, and financial achievements.
19:21 – Tesla Q3 financial and delivery highlights
Tesla reported strong performance across all regions with significant sequential growth in deliveries, driven largely by the excitement around the new Model Y variants released in Q1. The company highlighted its progress in operating a robotaxi service in Austin and Bay Area cities, emphasizing that their robotaxi vehicles blend seamlessly into the market without extra sensor equipment. They also noted ongoing efforts to expand coverage and regulatory approvals for Full Self-Driving (FSD) in multiple regions.
20:34 – Tesla continues to see progress in FSD adoption
Tesla continues to see progress in FSD adoption, although the paid FSD customer base remains around 12% of the fleet. They are working with regulators in China and AMIA to expand FSD availability. Financially, automotive revenues increased 29% sequentially in line with delivery growth, while automotive margins improved slightly due to better material costs and fixed cost absorption. The energy storage business achieved record deployments and margins but faces tariff-related headwinds, which Tesla is mitigating through its Shanghai mega factory supplying non-US demand.
22:21 – Tesla addressed tariff impacts
Tesla addressed tariff impacts exceeding $400 million for the quarter, split between automotive and energy businesses. Services and other segments showed sequential improvement, driven by gains in insurance and service center operations. Costs for the robotaxi program and other services like paid supercharging and used car sales are included here. Operating expenses increased, mainly due to restructuring aimed at cost reduction and efficiency improvements, legal expenses, and higher employee-related spend, especially in R&D linked to AI initiatives.
24:13 – Other income decreased
Other income decreased primarily due to lower gains from Bitcoin holdings and currency exchange movements. Tesla posted a record free cash flow of approximately $4 billion and ended the quarter with over $41 billion in cash and investments. Capital expenditures for the year are expected to be around $9 billion, with a significant increase projected in 2026 to support growth and AI initiatives like Optimus. The company concluded by emphasizing the challenges of bringing AI to the real world and expressing optimism about future benefits, thanking its team, customers, investors, and supporters.
25:23 – Robotaxi expansion and safety driver removal
The discussion begins with updates on Tesla’s Robotaxi metrics, including fleet size and operational progress. Tesla expects to remove safety drivers from large parts of Austin by the end of the year, emphasizing a cautious approach to deployment to avoid accidents. The company plans to operate Robotaxis in 8 to 10 metropolitan areas by year-end, contingent on regulatory approvals, with Nevada, Florida, and Arizona among the expected locations.
27:11 – Tesla continues to operate its Austin fleet without drivers
Tesla continues to operate its Austin fleet without drivers, having covered over a quarter million miles, while the Bay Area fleet with drivers has surpassed one million miles. Customer feedback is positive, and the fleet’s performance remains strong. Tesla’s supervised fleet has collectively logged 6 billion miles, marking a significant milestone. The company is on track to fully remove drivers from cars, starting with Austin. The next topic introduced is the demand and backlog for Tesla’s Mega Powerwall, solar, and energy storage systems amid the current AI boom.
28:24 – Mega Powerwall and solar demand growth
The demand remains strong heading into next year, driven by positive customer feedback on the Mega Mega Block product, which will start shipping from Houston. There is remarkable growth in AI and data center applications as hyperscale’s and utilities utilize the product to enhance grid reliability and relieve constraints. Residential solar demand in the US is surging due to policy changes, expected to continue into early 2026. A new solar lease product was introduced, and production of Tesla residential solar panels began in Buffalo, with shipments starting in Q1. These panels feature industry-leading aesthetics and shade performance, reinforcing the company’s commitment to US manufacturing.
29:57 – Challenges in Optimus robot manufacturing
Optimus robots are already capable of walking around Tesla’s Palo Alto engineering headquarters, assisting visitors by navigating them to specific locations. Developing a robotic hand as dexterous and capable as the human hand is extremely challenging due to the complex anatomy and function of fingers and forearm muscles.
30:59 – The forearm and hand of the Optimus robot
The forearm and hand of the Optimus robot present a more difficult engineering challenge than the rest of the robot due to the complexity of actuators and dexterity required. For a generalized useful robot, a highly capable hand combined with real-world AI is essential, along with the ability to scale production to relevant volumes.
32:12 – Tesla aims to manufacture Optimus robots at volumes
Tesla aims to manufacture Optimus robots at volumes comparable to or exceeding vehicle production, targeting potentially a million units per year. This manufacturing challenge is immense as there is currently no existing supply chain for humanoid robots, requiring Tesla to be highly vertically integrated and to produce many parts internally.
33:22 – Tesla’s unique position in manufacturing technology
Tesla’s unique position in manufacturing technology, real-world AI, and creating a truly dexterous hand sets it apart from other robot projects that lack these critical components. Achieving these goals requires immense effort, but Tesla’s strategy focuses on these key elements to realize a functional humanoid robot.
34:05 – A major concern expressed
A major concern expressed is maintaining strong influence over the deployment and control of a large Optimus robot workforce to prevent being ousted from Tesla. This concern underlies discussions about compensation and governance related to building a significant robot army. The conversation then shifts briefly to other topics, including Robotaxi expansion and semiconductor supply deals.
35:07 – Samsung chip deal and AI5 chip development
Elon Musk discusses Tesla’s chip manufacturing strategy, praising Samsung for producing the AI4 computer and explaining that both TSMC and Samsung will initially focus on manufacturing the new AI5 chip. He highlights his direct involvement with the AI5 chip design team and praises the design as exceptional.
36:34 – The AI5 chip is described as being 40 times more powerful
The AI5 chip is described as being 40 times more powerful than the AI4 chip due to Tesla’s deep integration of software and hardware design, allowing them to optimize the chip specifically for real-world applications. Tesla removed legacy components like the traditional GPU and image signal processor to streamline the chip’s functionality.
38:04 – Eliminating unnecessary components
By eliminating unnecessary components, the AI5 chip fits into half a reticle with room for critical elements like memory traces and ARM CPU cores. Both Samsung and TSMC fabs in the US will produce the chip, aiming for an oversupply to meet demands for cars, robots, and data centers.
39:41 – Tesla utilizes both AI4 chips and Nvidia hardware
Tesla utilizes both AI4 chips and Nvidia hardware in their data centers but emphasizes that Tesla’s chip design is simpler because it only needs to satisfy Tesla’s specific requirements. This focus on a single customer allows Tesla to remove complexity and optimize the chip more efficiently than Nvidia, which must cater to many customers.
40:50 – Tesla’s design approach
Tesla’s design approach reduces the complexity of chip interconnections, enabling radical simplicity and better performance. Elon predicts the AI5 chip will offer two to three times better performance per watt and up to ten times better performance per dollar compared to competitors, though mass production and scaling remain to be proven.
42:04 – Hardware 3 upgrades and FSD incentives
The discussion focuses on incentives for customers to trade in vehicles with Full Self-Driving (FSD) capability. Tesla has allowed FSD transfers to new vehicles and occasionally offered promotions to reward early adopters. The speaker emphasizes the importance of these customers and shares personal use of a hardware 3 vehicle with FSD daily. Additionally, after completing the V14 software release, Tesla plans to develop a lighter V14 version for hardware 3, expected around Q2 next year.
43:16 – Self-driving semi-truck production update
The discussion covers the production and deployment plans for autonomous semi-trucks, with the factory construction and equipment installation progressing on schedule. Validation trucks are already operating on roads, with larger production runs planned for late in the year and initial online builds expected early next year. The goal is to ramp up volume by the second quarter. While trains remain efficient for long-distance deliveries, autonomous semis are seen as a better solution for last-mile logistics and shorter distances. The team is also focused on advancing full self-driving technology for passenger vehicles, which is expected to be adaptable for semi-trucks once sufficient data is collected.
44:44 – Scaling production and Cybertruck outlook
Elon Musk discusses Tesla’s goal to expand vehicle production to an annualized rate of 3 million units within 24 months, contingent on supply chain capacity. The biggest production increase will come from the Cybertruck, optimized for full autonomy with no steering wheel or pedals, designed for comfortable rides rather than high performance. He emphasizes the focus on scaling production rapidly while maintaining margins and meeting strong demand.
47:41 – Elon highlights the key benefit of full self-driving
Elon highlights the key benefit of full self-driving, allowing passengers to text safely while the car drives itself which he believes will drive massive demand. He notes that autopilot significantly improves safety compared to distracted human drivers. Tesla’s software updates, including version 14.1, bring advanced AI features like enhanced reasoning and a world simulator that enables realistic reinforcement learning, dramatically improving the system’s capabilities.
50:13 – Elon describes Tesla’s AI roadmap
Elon describes Tesla’s AI roadmap, with the upcoming AI5 computer expected to be 10 times safer and 40 times more capable than AI4. He muses on the idea of cars potentially having excess intelligence and being ‘bored,’ suggesting the possibility of using the distributed computing power of millions of Tesla vehicles as a large-scale inference fleet. The conversation then shifts to the next question from Adam at Morgan Stanley, followed by Dan from Barclays.
52:18 – Tesla AI core competencies and future vision
Tesla started with zero core competencies but has grown into a company comprising multiple startups under one roof, initiated by Elon Musk. They developed products like battery packs for stationary and utility-scale use, and created the global Tesla Supercharger network, which other manufacturers in North America have adopted. Tesla’s chip design and AI software teams were built from scratch, reflecting the company’s approach of creating core competencies while advancing.
54:08 – The Tesla AI software team
The Tesla AI software team, formerly known as the autopilot team, was formed through active recruitment and innovation, with early hires like Ashok contributing significantly. The Optimus robot project exemplifies Tesla’s vision for AI-driven productivity, potentially achieving five times the output of a human annually by operating continuously without needing to recharge. This represents a step toward sustainable abundance where work becomes optional. Elon Musk emphasizes the limitless potential of embodied AI compared to traditional AI productivity enhancements.
55:26 – Tesla’s autopilot initiative began over a decade ago
Tesla’s autopilot initiative began over a decade ago, with Elon Musk describing the car as a ‘robot on wheels.’ Many engineers working on Optimus transitioned from vehicle engineering, bringing expertise in drive units and actuators. Tesla’s capability to scale manufacturing is rooted in this crossover. While the Optimus team includes experienced engineers, it also features many new graduates, highlighting a blend of fresh talent and institutional knowledge driving the project forward.
56:36 – The Optimus Engineering and manufacturing
The Optimus engineering and manufacturing teams work in an iterative loop to refine the robot’s design for functionality and manufacturability. Improvements have transformed Optimus from a rudimentary prototype to a more advanced version capable of humanlike movements, though earlier iterations like Optimus 2 were challenging to manufacture. This collaborative process balances innovation with practical production considerations.
57:48 – Optimus 2.5
Optimus 2.5 demonstrated advanced, humanlike movements such as kung fu, showcased publicly at the Tron premiere with Jared Leto. Despite clear robotic dimensions, many observers mistook Optimus for a human due to its fluid motions. The upcoming Optimus 3 is expected to bring significant improvements. Tesla holds regular, intensive review meetings to oversee the engineering and manufacturing progress of Optimus.
59:07 – Tesla’s AI efforts
Tesla’s AI efforts differ significantly from the large-scale AI model Grok developed by XAI, which aims at artificial general intelligence and requires massive compute resources that cannot be deployed in cars. Tesla’s AI models are much smaller and optimized for vehicle applications. While they approach AI from different angles, there is some complementarity; for example, Tesla uses Grok for voice recognition and generation in Optimus and integrates Grok’s voice capabilities in cars. Overall, Tesla and XAI pursue distinct AI strategies targeting different use cases.
1:02:01 – Robotaxi regulatory and deployment approach
The discussion focuses on the use of safety drivers in new markets for autonomous vehicles. Even if not required by regulators, the company prefers a cautious approach by having a safety driver or occupant during initial operations to monitor and address any unexpected challenges unique to a city, such as difficult intersections. This precaution helps prevent rare but critical incidents, typically lasting about three months until the system is deemed reliable enough to operate without a safety driver.
1:04:11 – FSD software safety and comfort improvements
The initial focus when releasing major new autopilot software updates is safety, with comfort improvements addressed afterward. Early software versions may feel jerky, so it is recommended that most users wait for version 14.2, where many comfort issues will be resolved.
1:04:41 – The conversation shifts
The conversation shifts briefly to differences between FSD14 and previous versions like FSD13, particularly regarding the feel and development paths for the Robotaxi technology.
1:05:16 – The development team is working extensively
The development team is working extensively on refining the real-world AI for autopilot, aiming to create a system that feels like a living entity. The roadmap for AI improvements is detailed and exciting, with leadership deeply involved in the process.
1:06:27 – The software for customer vehicles and Robotaxis
The software for customer vehicles and Robotaxis shares the majority of algorithms and architecture, with only minor feature differences such as parking options. Enhancements like AI reasoning are planned for release around versions 14.3 to 14.4, expected by the end of the year.
1:06:55 – The upcoming reasoning features
The upcoming reasoning features will enable the car to make smarter decisions, such as selecting parking spots based on likelihood and availability rather than just proximity to entrances. The AI will leverage 360-degree vision to outperform human abilities in tasks like finding empty parking spaces.
1:07:50 – A key challenge is fitting advanced reasoning
A key challenge is fitting advanced reasoning and AI capabilities into the car’s onboard AI4 computer, which must make real-time decisions. The team is focused on maximizing intelligence density within the hardware constraints.
1:08:19 – Tesla’s AI system
Tesla’s AI system is considered to have the highest intelligence density intelligence per gigabyte compared to competitors. This efficiency will provide significant advantages when transitioning to the more powerful AI5 computer, which will offer ten times the capability while maintaining the same intelligence density.
1:09:00 – Optimus production timeline and supply challenges
The discussion begins with addressing the challenges of hand dexterity in humanoid robots and the complexities of supply chain and vertical integration. The timeline for starting production next year is considered, with an emphasis on the fact that hardware design will not be fully frozen at production start. Iterations will continue as new difficulties arise, especially in making complex components like the new hand, which is described as an incredible engineering achievement.
1:09:59 – A production-intent prototype
A production-intent prototype for the robot is expected to be ready by Q1, around February or March. The plan involves building a production line capable of producing a million units of the Optimus robot, starting production towards the end of next year. The ramp-up to an annualized rate of one million units will be gradual, limited by the slowest or most problematic components. Future iterations include Optimus 4 targeting 10 million units and Optimus 5 aiming for 50 to 100 million units. The session ends with closing remarks and a thank you.
1:11:23 – Shareholder vote and Tesla leadership outlook
The discussion opens with an appeal to shareholders to support Elon Musk’s leadership through upcoming compensation proposals and board reelections. The speaker emphasizes the board’s integral role at Tesla and addresses the challenges of establishing voting control, noting that Tesla cannot implement super voting stock after going public, unlike companies such as Google and Meta that had this structure grandfathered in.
1:12:30 – There is concern over the influence
There is concern over the influence of proxy advisory firms like ISS and Glass Lewis, which often guide the voting decisions of passive index funds. The speaker describes these firms as corporate terrorists, criticizing their recommendations as sometimes harmful to the company’s future. This reliance by index funds on proxy advisors creates risks for corporate governance and shareholder interests.
1:13:42 – The core problem highlighted
The core problem highlighted is that index funds defer voting responsibility to proxy advisors, potentially leading to detrimental outcomes for publicly traded companies. The speaker praises the special committee for designing a shareholder-aligned compensation plan, urging shareholders to vote in favor of the plan and to reelect the three directors who are deeply involved in Tesla’s daily operations and governance.
1:14:57 – The session concludes
The session concludes with a final appeal to shareholders to follow the board’s recommendations and vote accordingly. The speakers thank the participants for their questions and express anticipation for the next quarterly meeting.
Further readings: Q3 2025 update letter
We’ve launched an official Tesla Owners Online Forum. Join early and claim your "Founding Fuel" badge.
