在2026年9月2日发布的一期播客中，埃隆·马斯克（Elon Musk）在一次G20采访中分享了“令人难以置信的预测”，对特斯拉（Tesla）和SpaceX的投资者具有重大影响。 **埃隆·马斯克的关键预测：** * **数字AI的经济影响：** 马斯克估计AI将使全球经济增长20%到30%，相当于每年20到30万亿美元的规模。他认为到明年底（2027年），AI将能够完成任何数字任务，只要它不需要手动塑形原子。他预计AI软件将达到“Stockfish（国际象棋引擎）水平的优秀”，这意味着它将在软件编写方面“碾压所有人类”，并在12到18个月内，在所有形式的工程和任何数字领域达到极其熟练的水平。 * **人形机器人的经济影响：** 马斯克预测，通用机器人AI将带来“全球经济的许多倍增长”，有可能使经济增长10倍或更多。他解释了一种递归效应：机器人将开始制造机器人，从而带来爆炸性增长。他保守估计，在10年内，将有远超10亿个人形机器人，每个机器人的生产力是人类的五倍。这意味着这10亿个人形机器人将比所有人类加起来更具生产力，他声称自己愿意“为此投入大量资金进行押注”。 * **电力/计算危机：** 马斯克强调了AI革命面临的“电力危机”。他指出明年（2027年）预计会出现严重的电力短缺，分析师的共识估计是，AI芯片的电力缺口至少达到15吉瓦。这是因为AI芯片产量每年增长40%到50%，而中国以外的电力供应每年仅增长10%到20%。他指出，谷歌（Google）和Anthropic已经从SpaceX租赁算力，而SpaceX通过建设自己的发电厂，已经能够建立AI基础设施。他认为这是各国为AI数据中心建设电力基础设施的一个机遇。 * **数字AI与物理AI的时间线：** 马斯克澄清说，任何物理事物总是比数字事物耗时更长，因为物理事物需要庞大的供应链和移动原子，而软件则可以轻松复制。他解释说，人形机器人的实用性是其AI软件、AI芯片和机电灵巧性的结合产物，所有这些都在呈指数级改进。 采访者指出，自Chat GPT推出以来，已有超过10亿人使用AI，并强调了马斯克关于人形机器人的“最高置信度预测”。采访者还证实了马斯克的“电力危机”预测，提到SpaceX正在建设一个用于涡轮叶片和叶片的工厂，以便利用天然气，因为目前制造商的等待时间已延长到2030年。他还重申了马斯克对数字AI的看好，并提及“crock bots”是最近的一项突破。主持人敦促观众思考，如果马斯克的预测哪怕接近正确，这将带来怎样的影响，特别是对于特斯拉（Optimus机器人）、SpaceX AI（计算业务）以及他们的“croc croc bot croc build croc”AI模型业务等公司而言。

In a podcast published on 2026-09-02, Elon Musk shared "mind-boggling predictions" during a G20 interview, with significant implications for Tesla and SpaceX investors. **Key Predictions from Elon Musk:** * **Digital AI's Economic Impact:** Musk estimates AI will increase the global economy by 20 to 30 percent, translating to an order of 20 to 30 trillion dollars per year. He believes that by the end of next year (2027), AI will be capable of doing anything digital that does not require shaping atoms by hand. He expects AI software to reach "Stockfish level good," meaning it will "crush all humans" in software writing and be extremely proficient in all forms of engineering and anything digital within 12 to 18 months. * **Humanoid Robotics' Economic Impact:** Musk predicts "many multiples of the global economy" from general-purpose robot AI, potentially increasing the economy by a factor of 10 or more. He explained a recursive effect where robots will start manufacturing robots, leading to explosive growth. He conservatively estimates that in 10 years, there will be well over a billion humanoid robots, with productivity per robot being five times that of a human. This means these billion humanoid robots will be more productive than all humans combined, a prediction he states he would be willing to "put serious money betting on." * **The Power/Compute Crisis:** Musk highlighted a "crisis of power" for the AI revolution. He stated there is an expected significant power shortfall next year (2027), with a consensus estimate among analysts of at least a 15 gigawatt shortfall of power for AI chips. This is because AI chip production is rising 40 to 50 percent a year, while power availability outside China is increasing at only 10 to 20 percent a year. He noted that Google and Anthropic are already leasing compute from SpaceX, which has been able to establish AI infrastructure by constructing its own power plants. He sees this as an opportunity for countries to build power infrastructure for AI data centers. * **Digital vs. Physical AI Timeline:** Musk clarified that anything physical always takes longer than digital due to the need for massive supply chains and moving atoms, as opposed to easily copied software. He explained that a humanoid robot's usefulness is a product of its AI software, AI chip, and electromechanical dexterity, all of which are improving exponentially. The interviewer noted that over a billion people are using AI since Chat GPT launched and underscored Musk's "highest confidence prediction" regarding humanoid robots. The interviewer also validated Musk's "crisis of power" prediction, mentioning that SpaceX is building a factory for turbine blades and veins to tap natural gas due to wait times from current manufacturers extending until 2030. He also reiterated Musk's bullishness on digital AI, referencing "crock bots" as a recent breakthrough. The host urged viewers to consider the implications if Musk's predictions are even close to right, particularly for companies like Tesla (Optimus robot), SpaceX AI (compute business), and their "croc croc bot croc build croc" AI model business.