In the second quarter, Tesla's total revenue reached $28.236 billion, marking a 26% year-over-year increase. This growth was propelled by higher vehicle deliveries, expansion in services and other business segments, and favorable foreign exchange impacts.

Following the earnings release, Tesla's senior management team, including CEO Elon Musk, CFO Vaibhav Taneja, Vice President of Supply Chain Karn Budhiraj, Vice President of Vehicle Engineering Lars Moravy, Director of Autopilot Software Ashok Elluswamy, General Counsel Brandon Ehrhart, and Head of Investor Relations Travis Axelrod, provided commentary on the results and addressed questions from analysts.

Initial Discussion Points

Ashok Elluswamy initiated the discussion by highlighting the performance of the Robotaxi program. He emphasized its exceptional safety record, noting that the autonomous taxi fleet has now driven over 380,000 miles without a human supervisor across six cities in two states, with no major incidents reported. He stated this achievement validates Tesla's AI-centric, camera-only approach to autonomous driving, countering historical beliefs that lidar, radar, and high-definition maps were essential for safety.

He also discussed the rapid scaling of the program, which began in Austin about a year ago. Since the start of this year, the fleet's weekly unsupervised mileage has been growing at a double-digit compound rate, a pace expected to continue. The effort required to expand to new cities is minimal and trending towards zero, aiming for statewide rather than city-by-city operations. This growth is supported by Tesla's world-class AI engineering team. The current Robotaxi fleet runs an early version of the v15 Full Self-Driving (FSD) software, which incorporates 40% of the planned improvements for that version. Completing v15 is expected to yield remarkable safety and performance.

Regarding the Cybercab, Elluswamy described it as an extraordinary product that instantly wins over anyone who tries it. Manufacturing targets have been aligned with the projected growth rate of unsupervised miles. The v15 architecture powering Model Y and other platforms will also be used in the Cybercab. In summary, the Robotaxi program is experiencing exponential growth with an impeccable safety record, and technological investments are paying off to sustain this trajectory.

On the topic of the Optimus humanoid robot, especially the third-generation model, the design goal is to match human form, function, and dexterity. This human-like design enables learning from observing humans perform a vast array of tasks, providing a wealth of training data. A dedicated data collection team also supplies high-quality demonstrations for fine-tuning. The second phase involves large numbers of Optimus robots practicing tasks in a dedicated facility, generating invaluable data for reinforcement learning and closing any remaining performance gaps. The AI strategy for Optimus is the same end-to-end approach used for FSD, developed by the same team, giving Tesla high confidence in its capabilities.

Analyst Q&A Highlights

Addressing a question about the Optimus supply chain, Elon Musk noted that supplier collaboration is progressing well, with major companies making significant investments. He specifically highlighted Samsung and Taiwan Semiconductor Manufacturing Company (TSMC) for their multi-billion dollar investments in U.S. fabs producing AI chips for robots and autonomous vehicles, and Amazon for expanding battery production. Karn Budhiraj added that new dedicated supply chains for robots, including memory chips and specialized components, are being built out separately from traditional automotive supply chains. Tesla will develop parts in-house if suitable suppliers cannot be found. Musk also thanked Micron Technology for allocating significant long-term memory chip capacity to Tesla amid a tight market.

Regarding Robotaxi regulation, Karn Budhiraj stated that progress is good at the federal level with the NHTSA, which has recognized vision-only systems in its updated safety standards. State-level regulations vary, with some, like New Jersey's, being overly restrictive. Tesla's preferred regulatory model sets safety outcome goals but allows companies to determine the technological path to achieve them. The company relies on real-world safety data to persuade regulators and the public.

When asked about key milestones for scaling Robotaxi, Elon Musk emphasized the goal of vertical integration for the business. The primary constraint is achieving what he termed the "journey to five nines" of reliability (99.999%). Once this level of reliability is reached, he believes demand will far outstrip service capacity.

On the potential merger of Tesla and SpaceX, Musk stated that while collaboration is deepening across many areas, including projects like Terafab and Digital Optimus, a merger cannot be discussed on an earnings call and would require proper procedures. General Counsel Brandon Ehrhart added that the companies benefit from their partnership, which was deepened earlier this year through an investment and framework agreement.

Musk elaborated on the integration of Starlink into vehicles, especially the Cybercab, noting its critical role in providing ubiquitous connectivity for autonomous taxis, which cannot afford dead zones. He also mentioned the potential for vehicles to act as connectivity relays. Vaibhav Taneja added that the in-car experience with a large screen becomes far more valuable when passengers don't need to focus on driving.

Explaining the strategy of expanding Robotaxi service to multiple cities with small fleets rather than concentrating in one, Ashok Elluswamy said it demonstrates the system's broad generality and requires little incremental effort per city. The exponential growth is in its early stages. He also noted that because Robotaxis run almost continuously, even a small fleet can accumulate high mileage. Vaibhav Taneja added that scaling across cities helps identify and resolve operational bottlenecks in a controlled manner before massive deployment. Lars Moravy cited varying local traffic regulations as another reason for city-by-city expansion.

On federal regulations, Lars Moravy expressed a good relationship with the NHTSA, which he sees as trying to get ahead of public expectations. He does not believe the removal of mandatory steering wheel rules will be a core bottleneck for Cybercab scaling.

Regarding the roadmap for an autonomous Tesla Semi truck, Musk acknowledged the severe driver shortage and the significant market need. However, with Semi production still low, the focus for the next ~6 months remains on achieving full self-driving for the high-volume Model 3/Y and Cybercab. Work on an autonomous Semi is expected to ramp up next year as its production scales.

On the semiconductor strategy for Optimus, Musk noted that Optimus 3 uses many custom Tesla-designed components manufactured by suppliers. For the higher-volume Optimus 4, targeted for production in Austin, the supply chain will be more vertically integrated, with Tesla likely doing much of the printed circuit board work internally to achieve a targeted order-of-magnitude increase in annual production.

Regarding hardware upgrades for older vehicles to run newer FSD software like v15, Musk said it makes economic sense to upgrade cars designed for cameras. Tesla plans to offer an upgraded AI board, an intermediate improvement over Hardware 4, around mid-next year, followed by Hardware 5. He praised the Tesla chip team's progress, stating their AI chip will be the world's best edge computing chip.

On capital expenditure (CapEx) strategy, Musk stated the goal is to spend as fast as possible without waste, balancing capital efficiency with speed to maximize net present value. He expressed satisfaction with the unprecedented scale and pace of Tesla's industrial expansion across multiple fronts. Vaibhav Taneja added that all CapEx is directed toward productive assets like factory expansions for Optimus, Cybercab, LFP batteries, semiconductors, and solar manufacturing. Musk emphasized that the market underestimates the potential for scaling U.S. solar manufacturing, which Tesla plans to increase by an order of magnitude.

Finally, discussing the energy storage business, Musk highlighted applications beyond peak shaving, such as grid balancing to support wind and solar power. He identified power constraints as a major future challenge for AI compute, where advanced power electronics are needed to smooth massive, rapid power fluctuations during training runs—a key use case for Tesla's Megapack. He argued that batteries could potentially double U.S. energy output by enabling more efficient use of existing generation capacity, leading to very high future demand for Megapack.