Tesla’s Q3 2024 call: cheaper cars, cash flow and autonomy forecasts

Musk projected 2025 vehicle growth and a driverless service, while Tesla reported $6.3 billion in quarterly operating cash flow.

EMElon Musk登壇者 · TeslaVTVaibhav Taneja登壇者 · Tesla
チャンネル
Humanoid Robots Guide To The Galaxy
公開日
2024年10月23日
動画
1:05:05
読了時間
7 分
原文
英語
YouTube で開く
スクロールすると動画は右下に小さくなって再生を続けます。字幕は閲覧言語に合わせて切り替わります。
閲覧言語2 言語で読めます
この記事はまだ日本語に翻訳されていないため、英語の原文を表示しています。
要約

30 秒でわかる

On Tesla’s October 2024 earnings call, Elon Musk forecast 20% to 30% vehicle growth in 2025 and said lower-cost models were planned for the first half of that year. He also projected a driverless ride service, subject to regulatory approvals, and Cybercab volume production in 2026. CFO Vaibhav Taneja reported $6.3 billion of quarterly operating cash flow and said software features contributed $326 million in revenue. Those autonomy and production timelines were forecasts, not completed milestones.

  1. 1Musk forecasts 20% to 30% vehicle growth in 2025He links the projection to lower-cost models and autonomy, while noting that major external events could change the outcome.
  2. 2A human-safety crossover is an internal estimate, not a demonstrated resultMusk says Tesla estimated a critical-intervention threshold could be reached around the second quarter of 2025; the company’s supervised product requires an attentive driver.
  3. 3Tesla reports $6.3 billion in quarterly operating cash flowTaneja says financing incentives weighed on average selling prices and specified FSD feature releases contributed $326 million in revenue.
  4. 4Energy gross margin exceeds 30%, with project mix driving variationEnergy deployments fell from the preceding quarter, but Taneja expected a fourth-quarter increase and annual growth.
  5. 5Tesla is unsure whether older hardware can support unsupervised FSDMusk says the company would upgrade eligible FSD customers at no charge if Hardware 3 proves insufficient.

ポイント

  1. In October 2024, Musk forecast 20% to 30% vehicle growth in 2025 and planned lower-cost models for the first half of the year.
  2. Tesla reported about $6.3 billion in third-quarter operating cash flow; $326 million in revenue came from specified FSD feature releases.
  3. Musk projected driverless ride-hailing and Cybercab production, while acknowledging approval and technical uncertainty.
  4. Tesla’s FSD (Supervised) requires an attentive driver; the company’s future autonomy claims are projections.
  5. Taneja said energy gross margin exceeded 30% in the quarter but could vary with project mix.
全文

チャプターごとに整えた全文を読む

文字起こしのみ
01

Vehicle growth, battery costs and autonomy plans

Musk reviews quarterly deliveries and makes forecasts for lower-cost models, Cybercab and FSD. The ride-hailing test he describes used safety drivers.

Elon Musk opens Tesla's October 2024 earnings call by saying the company remained profitable and achieved record third-quarter deliveries in a difficult automotive market. He says Tesla produced its seven-millionth vehicle the day before the call.

Musk says demand for Megapack and Powerwall energy storage is growing. He recaps the October robotaxi event, describing Cybercabs and Model Ys carrying visitors around a controlled venue and humanoid robots interacting with the crowd.

Musk says Tesla remains on track to introduce more affordable models in the first half of 2025. In a forecast made on the call, he estimates vehicle growth of 20% to 30% in 2025, subject to major external shocks.

He expects Cybercab volume production in 2026 and offers an eventual target of at least two million units a year, potentially more. These figures are management projections, not results achieved at the time of the call.

Musk says Tesla's 4680 battery cell is approaching a competitive delivered cost in North America after incentives and duties. Even if internal production improves, he says Tesla will continue to buy many cells from suppliers as vehicle and storage production grows.

On Full Self-Driving, or FSD, Musk describes software updates and expects version 13 to improve the distance between interventions by five to six times compared with version 12.5. He and an executive expect further improvements as training, data and computing scale.

Musk says Tesla's internal estimate in October 2024 was that its system could exceed human performance on a critical-intervention measure around the second quarter of 2025, while acknowledging that timing could slip. This is a company forecast, not an independent finding.

Musk says Tesla planned further FSD trials to encourage owners to try updated software and predicts that existing vehicles could eventually operate without supervision. Tesla's [FSD (Supervised) manual](https://www.tesla.com/ownersmanual/model3/en_us/GUID-2CB60804-9CEA-4F4B-8B04-09B991368DC5.html) requires a fully attentive driver ready to act; the product name does not make the car autonomous.

Tesla executives say employees in the San Francisco Bay Area were testing a ride-hailing application with a safety driver. They describe existing features, including synchronized driver profiles and mobile navigation, as building blocks for a future service.

In 2024 Musk says he expects a public ride-hailing service in Texas and possibly California in 2025, subject to the relevant approvals. He characterizes the prospective service as a change in Tesla's business beyond manufacturing vehicles and energy equipment.

Musk cites a Tesla vehicle-safety report comparing Autopilot crash rates with a U.S. average. Those figures describe different driving populations and conditions and do not by themselves establish a tenfold causal safety benefit.

He says assessing successive driving models becomes harder when errors take longer to find. Tesla is expanding AI training capacity, but executives say identifying and validating differences between versions has become a constraint.

02

Optimus, storage and the scale of energy

Musk outlines robot and energy-storage ambitions, including the planned Shanghai Megapack factory.

Musk discusses a redesigned Optimus hand and forearm with more degrees of freedom and improved tactile sensing. He argues that AI and high-volume manufacturing are necessary to scale humanoid robots, and predicts Optimus could become an exceptionally valuable product.

For stationary storage, Musk says the Lathrop Megapack factory reached a production rate equivalent to about 40 gigawatt-hours a year. He says a Shanghai plant is expected to start in the first quarter of 2025 with a 20-gigawatt-hour annual run rate.

Musk argues that a sustainable energy system built around generation, batteries and electric transport is physically possible with available materials. He uses the Kardashev scale—a way of discussing how much energy a civilization can harness—to emphasize the scale of solar energy.

03

Cash flow, margins and investment

Taneja discusses financing incentives, software revenue, energy margins and AI infrastructure spending.

Chief financial officer Vaibhav Taneja reports approximately $6.3 billion in operating cash flow for the quarter. He says automotive revenue increased as deliveries grew, although financing incentives reduced average selling prices.

Taneja says FSD features, including Cybertruck availability and Actually Smart Summon, contributed $326 million in quarterly revenue. He also notes more than $2 billion in regulatory-credit revenue year to date and says Tesla wants to grow sales without accumulating inventory.

Taneja attributes improved automotive margins to software revenue, higher volume, commodity costs and more localized deliveries, but says sustaining those margins in the fourth quarter will be challenging. Musk adds that lower interest rates would help monthly vehicle payments.

Energy deployments fell sequentially in the third quarter, Taneja says, but he expects them to rise in the fourth and more than double for the year. Energy gross margin exceeded 30% in the quarter, though he cautions that margins vary with the project mix.

He says Tesla reached its lowest cost per vehicle in the quarter and that services benefited from repairs, parts and the growing Supercharger network. Operating expenses declined partly because of restructuring, with some savings offset by AI-related costs.

Taneja says Tesla is deploying a GPU cluster in Austin and expects full-year capital expenditures above $11 billion. He says bringing the autonomy and robotics vision to market will take time and require navigation of different regulatory regimes.

04

Affordable vehicles, service and the Semi

Executives discuss lower-cost cars, Tesla’s autonomous strategy, service capacity and plans for electric truck production.

Asked about an affordable vehicle, executives reaffirm a first-half 2025 plan. They distinguish lower-cost personally owned models from the proposed robotaxi and discuss a sub-$30,000 price after incentives.

When asked about a conventional $25,000 non-robotaxi car, Musk says Tesla is prioritizing autonomous vehicles and considers a conventional version inconsistent with that strategy. He says the Cybercab is designed without a steering wheel or pedals and could be available for individual purchase.

On service wait times, executives describe fixing problems at the factory, diagnosing vehicles remotely and expanding service locations. They say larger centers can use dedicated lanes for different repairs, while software can prepare parts and technicians before a car arrives.

Tesla executives say construction of the Semi factory near Reno is progressing, with pilot builds planned in the second half of 2025 and production in the first half of 2026. They say adoption will depend on customers, while Musk predicts strong demand if total cost per mile is competitive.

Executives cite experience with Tesla's fleet and Pepsi as they discuss the Semi's performance and driver acceptance. They say trucks being built have cameras and hardware intended to support FSD, with software training still underway.

05

Regulation, manufacturing and older hardware

The team answers questions on driverless approvals, vehicle cost reduction and whether Hardware 3 can support unsupervised FSD.

Asked about X and Grok in cars, Musk says Tesla plans to expand in-car software, browsing and media. He imagines passengers watching films or using other applications in a future autonomous vehicle.

An investor asks whether regulators have approved unsupervised FSD across California and Texas. Musk says they have not: states have different processes, and he expects California approval but acknowledges that it is not fully within Tesla's control.

Executives explain that federal vehicle standards and state deployment rules play different roles. Musk argues for a national approval process for autonomous vehicles rather than navigating fifty separate state systems.

On lower-cost cars, Musk describes manufacturing cost reduction as thousands of small changes rather than one breakthrough. He gives a simplified example: saving an average of 50 cents across 10,000 parts would reduce cost by $5,000.

Executives say Cybercab's design and production system are being developed together, including decisions about parts, shipping and assembly. They also list 2025 work on storage, charging, battery cells and lithium refining.

Musk thanks people who placed deposits on the delayed next-generation Roadster. He says the company is still working toward a final design but has prioritized projects it sees as more central to its mission.

Asked whether older Hardware 3 vehicles can achieve unsupervised FSD, an executive says the team develops first on more capable Hardware 4 and then adapts the software. Musk says Tesla is not certain Hardware 3 will reach the required safety level.

If Hardware 3 proves insufficient, Musk says Tesla would upgrade the relevant computers for customers who purchased FSD at no charge. He says the company designed the system with an upgrade in mind, while emphasizing that the outcome was not yet known.

06

AI training, safety drivers and xAI

Analysts ask about compute, state deployment rules and the relationship between Tesla and Musk’s separate AI company.

Analyst Pierre Ferragu asks whether additional computing is used to train larger models or train existing ones faster, and how Tesla would deploy a ride service. Musk says driving requires processing large amounts of camera data on a relatively power-limited computer inside the car.

Musk says more training can help a small in-car computer perform efficiently. Executives add that selecting useful data and validating improved models take substantial work; they say Tesla intends to follow each state's requirements for safety drivers and deployment.

Morgan Stanley analyst Adam Jonas asks whether Musk's AI company xAI competes with Tesla for talent, technology and his time. Musk says xAI has helped with large training runs, but argues that its general-purpose AI work and Tesla's vehicle and robot systems address different problems.

Musk contrasts large server installations with the limited power available for computing in a car. He says xAI aims to develop a truth-seeking general AI, while Tesla focuses on real-world systems; the call concludes after this answer.