股东会·中文精读

观看:马斯克在 2021 Tesla 股东大会完整演讲

Watch Elon Musk at 2021 Tesla Shareholder Meeting (full presentation)

中文精读由 AI 基于原始视频与文稿改写,非官方翻译,可能存在偏差或错误;表述以原视频为准。说明

2021 年 10 月,特斯拉股东大会在得州奥斯汀的超级工厂里开。马斯克上台先感谢团队,第二件事是翻旧账。

他说几年前自己宣称特斯拉能做到每周生产 5000 辆 Model 3,差点因为这句话被抓起来。

"我当时真的差点因为宣称我们能做每周 5000 辆而被捕。现在说起来大家都在笑。"(马斯克)

台下确实笑了。那个在 2018 年需要反复辩护、被媒体逐条质疑的数字,放到 2021 年已经不值一提。特斯拉当年的交付量创了纪录,而这条增长曲线,马斯克认为可能是有史以来大型制造物体里最快的一条。

这场发布会信息密度很高。交付、毛利率、电池、自动驾驶授权、工厂选址、AI,一个接一个往外抛。下面按主题拆开讲。

每周 5000 辆,曾经是个会让人惹上麻烦的数字

马斯克给了一张交付曲线图,然后说了一句很少见他说的自我评价:这可能是任何大型制造物体增长最快的一次,至少是最快的之一。

他没有把功劳全揽在自己身上,反复强调的是"如果芯片能拿到,我们就能做到"。

2021 年全球缺芯,汽车行业普遍减产,特斯拉是少数还在往上走的。马斯克的表述很实在:未来能不能维持,很大程度取决于供应商的挑战。他希望芯片短缺尽快缓解,同时表示自己有信心在相当长一段时间里维持 50% 以上的增长。

这个承诺值得拆开看。50% 的复合增长意味着四年翻五倍,对一家已经年销接近百万辆的公司,难点从来都落在物理世界:产线节拍、供应链、电芯产能、码头和铁路。马斯克在台上几乎没谈需求,谈的全是供给,这本身就说明了瓶颈在哪。

他还顺手把 Model 3 的成绩单亮了出来:全球最畅销的高端车,不分能源类型。

这句话放在 2018 年会被人当笑话。当年为了让 Model 3 的产能爬出"生产地狱",他睡在工厂地板上,接受采访时候眼眶发红。四年后,同一款车成了全球高端车销量第一。

Model Y 的时间表:2022 年看营收,2023 年看数量

接下来是全场最具体的一组预测。

马斯克说 Model Y 会超过 Model 3,成为全球任何类别里最畅销的车型。他给了两个时间点:2022 年按营收算最畅销,2023 年按数量算最畅销。

这个排序有讲究。按营收第一,说明单车价格和销量的乘积领先;按数量第一,意味着要超越卡罗拉、RAV4 这类年销百万级的家用车。从营收第一走到数量第一,中间隔着一次产能释放。

所以他紧接着说了前提:奥斯汀和柏林两座工厂必须上线,并且达到量产。

这句话听着像客套,放在当时的语境里是全场最硬的约束。新工厂从第一台车下线到周产几千台,通常要一年以上。马斯克把 2023 年这个数字说出口,等于把整个制造团队的时间表钉在了墙上。

Model Y 和 Model 3 共用大量零部件,这是它能快速上量的基础。平台复用带来的成本优势,在传统车企里往往需要几代车型才攒得出来。

从 2018 年那种日子爬出来,他不想再回去

讲到现金流,马斯克的语气变了一下。

他说 2017 和 2018 年那段时间,情况"至少可以说有点悬"。原话是"不想再回到那里去"。

2018 年的特斯拉确实站在悬崖边。Model 3 产能爬坡烧掉大量现金,公司一度离破产只有几周距离,马斯克后来承认当时正在和多家银行谈紧急融资。那段经历在他身上留下了明显的痕迹,此后每次谈增长,他都会先谈现金。

2021 年的情况完全反过来了。自由现金流持续为正,马斯克预计会继续强劲。他的逻辑是一个乘法:单位销量乘以自动驾驶,再乘以工厂效率的提升。

这个公式里,销量是基数,工厂效率决定成本,自动驾驶充当乘数。三者叠加,才有他口中"非常好看的数字"。

他还讲了一个判断:所有制造商最终都会造电动车,然后所有制造商最终都会造自动驾驶车。他描述的是一个行业终局,没有预测哪家公司会赢。传统车企的转型也就无从回避,只剩时间早晚的问题。

自动驾驶要授权出去,理由跟赚钱无关

顺着上面那个判断,马斯克说了一句在汽车行业相当罕见的话:特斯拉对授权自动驾驶持开放态度。

他给的第一个理由是安全。

"自动驾驶会挽救大量生命、避免大量伤害,这项技术不该只留给我们自己。如果有其他厂商想用,授权给他们,在道德上是正确的。"(马斯克)

这段话可以说得很漂亮,背后的商业逻辑同样清晰。

特斯拉真正的护城河是数据规模和算力,算法本身排在其后。授权给其他厂商,等于把更多车队接入自己的数据循环,同时把竞争对手变成客户。安卓当年就是这么做的,代价是放弃了授权费之外的全部硬件利润,收益是拿到了整个生态。

区别在于,汽车的安全责任比手机重得多。一套自动驾驶系统如果装到别的品牌车上出了事故,责任怎么划分,监管怎么认定,这些马斯克在台上没有展开。

他只用一句话带过:这项技术会成为救命的工具,所以不该藏着。

一边涨价,一边把成本往下压

财务数据里有一组看起来矛盾的数字。

特斯拉的平均售价在明显下降,因为 Model 3 和 Model Y 比 Model S 和 Model X 便宜得多。通常售价下降会压低毛利率,但特斯拉的毛利率反而在上升。

马斯克自己也承认这很难:"把平均价格降下来,同时把毛利率提上去,是非常困难的事,但我们做到了。"

原因不神秘。工厂产能利用率提高,单车分摊的固定成本下降;零部件本地化和平台复用降低物料成本;上海工厂的高效率把制造成本压到了行业低位。这几件事叠加,才让降价和毛利率上升同时成立。

但他也承认了另一面:供应链成本压力很大,特斯拉不得不涨价,至少是暂时的。他希望未来能把价格降回去,让车尽可能便宜。

还有一句很实在的吐槽:为了保交付,特斯拉花在空运零部件上的钱"多得离谱"。这笔钱在财报里体现为费用,在交付量上体现为数字。2021 年很多车企选择少交付,特斯拉选择花钱换交付。

目标是让车尽可能便宜。这句话他重复了不止一次。

对供应商说的话:你能造多少,我买多少

进入电池环节,马斯克先解释了自己那件印着字的文化衫,是电池日的周边,设计得很隐晦。

然后他说了正题:特斯拉需要大量电池,多到需要供应商和自产两条腿走路。

有供应商直接问他,你们是不是打算把我们干掉了。他的回答很直接:完全没有这回事。

"你们想卖多少就卖多少,只要价格合适,我们全买,没有上限。有供应商问能不能增加一倍,我说没问题。"(马斯克)

这话听着像客套,放在 2021 年的背景下是真实的焦虑。那一年所有车企都在抢电芯,谁拿到电芯谁就能交付。特斯拉的策略是把供应商的产能尽可能锁死,同时自己建厂补缺口。

他还讲了一个分配原则:电芯优先给车,多余的部分给 Powerwall 和 Megapack。

这个优先级说明了当时储能在特斯拉内部的地位。车是现金流主力,储能是战略储备。但从长期看,他认为固定储能的需求至少会和车一样高。

可持续能源只需要三样东西,加一块一平方英里的电池

这是全场最有"马斯克味"的一段。

他把可持续能源的未来拆成三个支柱:太阳能和风能作为发电主力,固定电池储能,电动交通。三样齐了,可持续能源的未来就成立了,只要太阳还照常升起。

有人问他怎么看核聚变。他的回答把全场逗笑了。

"我喜欢聚变这个想法,而且我认为它完全可行。但天上本来就有一个巨大的聚变反应堆,每天早上准时出现,零维护。我觉得这笔买卖挺划算。"(马斯克)

他接着补了一句:捕捉太阳光里的一小部分,就能给整个地球供电。

然后给了一个具体到让人意外的数字:大约几百英里见方的太阳能板,就能给整个美国供电。电池只需要一平方英里。

这两个数字第一次听会觉得不真实,算一下就对上了。美国年发电量约 4 万亿千瓦时,光伏每平方米装机约 200 瓦,年发电约 300 度。理论上需要的面积在几万平方公里量级,也就是两百多公里见方。卡住这件事的从来是电网、储能和审批,面积本身排在很后面。

储能之所以必要,是因为风光都是间歇的。风不会一直吹,太阳不会一直照。电力系统需要缓冲,电池就是这个缓冲。马斯克把这件事说得像常识:你需要电池来给电力做缓冲,就这么简单。

特斯拉的价值,是"提前了多少年"

有人问特斯拉的长期价值怎么衡量。马斯克没有回答销量,没有提利润率,也没提市值。

他说,衡量特斯拉价值的根本方式,是它把可持续能源的到来提前了多少年。

"特斯拉的根本价值,在于我们加速可持续能源多少年。提前的年数越多,价值越大。"(马斯克)

这个定义解释了特斯拉很多看起来不划算的动作:开放专利、授权自动驾驶、自己建电池厂、做储能。站在季度利润的角度,这些都不是最优解;站在"加速多少年"的角度,它们全都说得通。

他顺势聊到碳税。有人说碳税会利好特斯拉,他承认没错,但补了一句:碳税也会伤到 SpaceX。

"有人说碳税会让特斯拉受益。是的。但它也会伤害 SpaceX。那碳税到底该不该有?它确实是必要的。"(马斯克)

这段自嘲里有个真实立场。碳税把化石燃料的外部成本内部化,对所有烧油的公司都是负担,对做电动车和太阳能的是顺风。他愿意承认自己受益,也愿意承认自己的另一家公司受损。

他还提到美国的平均生命周期排放会随着电网变清洁、车辆电动化继续下降。这件事没有捷径,只能一年一年往前挪。

上海工厂:质量最好,成本最低,戏最少

讲到工厂,马斯克的语气明显兴奋起来。

他说上海工厂的表现令人难以置信,产量已经超过弗里蒙特。然后他停下来,专门请全场给特斯拉中国团队鼓掌。

他的评价只有三个"最":质量最好、成本最低、戏剧性最少。

第三点最耐人寻味。弗里蒙特工厂的历史充满波折,产能爬坡期间出过各种事故、劳资摩擦和媒体风波。上海工厂从奠基到投产只用了不到一年,之后一路平稳扩产。对一个制造业老板,没有意外比任何单项指标都珍贵。

弗里蒙特这边,特斯拉计划把产量再提高 50%,Model S 和 Model X 仍然只在弗里蒙特生产。

他解释了一个布局逻辑:高销量的车型应该在离消费者最近的大陆生产。把车从上海运到欧洲,几千英里的海运既不经济也不环保。所以有了柏林,有了奥斯汀。

奥斯汀超级工厂和柏林工厂都在建,两座工厂内部都会包含电芯生产。马斯克用一句话描述这种模式:原材料进去,整车出来。

他甚至开了个玩笑,说这些工厂的规模要用"五角大楼"作单位来计量。听着夸张,但奥斯汀工厂的占地面积确实超过一千万平方英尺。

电池是高品位矿石,不是用完就扔的垃圾

有人关心电池的环保问题,马斯克的回答从化学讲起。

他说二氧化碳是极其稳定的分子。火星大气主要成分是二氧化碳,几十亿年都没变过。所以不用担心电动车排放的二氧化碳会怎样,它本来就在自然界大量存在。

甲烷倒是有人担心。他说不用太担心,甲烷很快会分解成二氧化碳,它不是一个稳定的分子。

然后是回收。他给了一个很好用的比喻:电池就是高品位矿石。

锂、镍、钴这些材料,可以从岩石里挖,也可以从旧电池里提取。从电池里提取的品位更高,成本更低,能耗更小。回收电池的经济账本身就算得过来,不需要靠补贴驱动。

特斯拉已经开始做电池回收,他预计其他电动车公司也会跟进,因为这件事本身能赚钱。

这段话把电池回收从"环保义务"重新定义成了"采矿替代"。一旦这么看,回收的经济属性就从成本项翻成了资源项。这个视角转换值得记住。

那场大雪,和一间没电的房子

马斯克讲了一段亲身经历。

2021 年初得州遭遇极端寒潮,电网崩溃,几百万人断电。他当时就在奥斯汀,住在一间没有电、没有灯、没有暖气、没有网络的房子里。想去超市买点东西,超市里也没有食物。这种情况持续了好几天。

然后他说了一句很有画面感的话:如果那间房子装了太阳能板和 Powerwall,车会有灯有电,加上星链,网络也不会断。

他顺口开了个玩笑,说这些正好是末日准备者需要的东西,真到了世界末日能派上用场。

这段闲聊其实是储能业务最好的广告。极端天气越来越频繁,电网的脆弱性暴露得越来越彻底。一个家庭如果同时有光伏、储能和电动车,就具备了最基本的电力韧性。这个需求在 2021 年之后涨得很快,原因很清楚。

他还提到了工厂安全。特斯拉的工伤率比行业平均好 18%,目标是建成地球上最安全的工厂。他承认新产线投产期间安全数据会变差,等进入稳定状态、工人熟悉流程、问题被逐个解决之后,受伤率会自然下降。

特斯拉是家软件公司,只是顺便造车

讲到 AI Day,马斯克说那场活动的意义在于改变一个认知。

人们习惯把特斯拉看成汽车公司。车确实造得不错,但特斯拉的 AI 部分长期被低估。他的原话是:特斯拉既是硬件公司,也是软件公司,两者分量相当。

证据是芯片。特斯拉自己设计了全自动驾驶的推理计算机,还在设计一台训练计算机,他认为这台机器会成为世界上最高效的神经网络训练计算机,而且领先幅度不小。

一家车企自己流片做芯片,在行业内是罕见的。背后的原因是通用 GPU 太贵、太耗电,而自动驾驶对推理延迟和功耗的要求极其苛刻。买不到合适的,就只能自己做。

AI Day 之后,特斯拉的求职申请曲线明显上翘。马斯克专门放了这张图,说效果超出预期,团队干得很漂亮。

这件事的启示是:一家公司对外讲什么故事,会直接改变它能招到什么人。在 AI Day 之前,顶尖 AI 研究者不会把特斯拉当成首选;之后,情况变了。

他对 AI 的心情很复杂

谈到 AI,马斯克的语气收了一下。

他说自己对 AI 有混合的想法。他认为 AI 会是未来极其重要的一部分,自动驾驶只是其中一个功能;同时他也警惕 AI 变成危险。

但他没有停在警告上。

"无论如何它都会发生。如果我们在其中,至少可以尽最大努力,确保它是一个好的 AI。"(马斯克)

这段话是他多年来对 AI 立场的浓缩版。他一边公开呼吁监管,一边自己下场做,包括后来的 xAI。听起来矛盾,逻辑其实一致:这件事挡不住,那就在牌桌上争取一点影响方向的机会。

他还提到特斯拉努力在每件事上都做正确的事。

"我们不总是成功。但如果你在找一家真的在努力做正确事情的公司,那就是特斯拉。"(马斯克)

这句话放在股东大会上说,真正的听众是员工和应聘者。一家公司对外宣称的价值观,最终会筛选出相信它的人。

总部搬去得州,加州还要扩产 50%

最后,他宣布了那条当天被转发最多的消息:特斯拉总部从加州帕洛阿托搬到得州奥斯汀。

掌声之后,马斯克马上补了一句:这不代表特斯拉离开加州。公司打算把弗里蒙特工厂和内华达超级工厂的产量各提高 50%。

他解释搬家的原因,用的比喻很生动。弗里蒙特工厂已经塞满了,像沙丁鱼罐头。刚接手那会儿,工厂大得像个巨人在穿小孩的鞋,现在反过来了,是小孩的脚硬要塞进巨人的鞋里,怎么摆都摆不下。

还有更现实的理由:湾区的房价太高,很多员工要长途通勤才能上班。工厂可以往上盖,人的通勤半径有上限。

奥斯汀这边,新工厂离机场五分钟,离市中心十五分钟。工厂挨着科罗拉多河,他承诺要建一座"生态天堂"。

这个选址逻辑值得琢磨。制造业选址看三件事:地、人、物流。得州地便宜、税低、监管相对宽松,奥斯汀又是过去十年美国人口流入最快的城市之一,工程师供给充足。加州继续做研发和高端车型,得州做高量产车型,分工很清楚。

把这场发布会当成一份战略说明书

整场会议听下来,马斯克其实只讲了一件事:特斯拉的增长受什么约束,以及他打算怎么解开这些约束。

约束有四个:芯片、电芯、工厂产能,以及湾区的地理天花板。对应四个动作:锁定供应商、自建电芯产线、三地同时建厂、总部搬迁。

所有动作指向同一个目标,就是他自己定义的那个价值标准:把可持续能源的到来提前多少年。

这个标准不能塞进财报,但它解释了这家公司几乎所有看起来不合理的决定。

  • 看一家制造业公司的增长承诺,先找它的物理瓶颈。马斯克整场没谈订单,只谈芯片和电芯,说明供给端才是卡口。判断任何产能扩张故事,都该先问一句:卡在哪一种物料上?
  • 盯住储能这条线。马斯克明确说固定储能的长期需求至少和车持平,而当时的产能分配是"车优先,多余的给储能"。供给被压着的市场,通常藏着机会。
  • 评估降价和毛利率能否同时改善。如果一家公司平均售价在降而毛利率在升,去查它的工厂利用率和零部件自制率,这两项才是原因。
  • 把"授权"当成一个信号。愿意把核心技术授权出去的公司,赌的是生态和数据规模,单点技术壁垒已经让位。这类公司的长期价值要按平台算,不按产品算。
  • 家里如果要装光伏,把储能一起算进预算。马斯克讲的那场得州大雪是真实的电力韧性案例,跟营销故事无关。断电时能维持基本用电,这个价值在极端天气变多的年份会越来越具体。
  • 电池回收值得单独跟踪。把旧电池当高品位矿石看,这个行业的经济模型就不依赖补贴,靠的是材料价差。
  • 看一家公司怎么讲自己,就知道它能招到谁。AI Day 之后特斯拉求职申请激增,说明叙事会改变人才流向。评估一家公司的技术前景,看它的招聘曲线比看它的发布会更准。
YouTube 平台字幕(自动或人工),可能有识别错误,人名与数字请以视频为准。已按语句合并成段落,并去掉 um、uh 等口头填充词。 文稿来源:YouTube(自动/人工字幕)。

0:04[Applause] all right welcome to the annual shareholder meeting coming to you live from austin texas so we've had a fantastic year thanks to the great work of the tesla team i'd just like to start off by just thanking the tesla team for the incredible work over the past year to get where we are well then

0:37so we have record vehicle deliveries as we've already reported this but you can see that there's a pretty i mean i think this might be the fastest that any large manufactured object has grown like yeah certainly one of the fastest perhaps the fastest and it looks like we we have a good chance of maintaining that into the future really dependent on

1:08supplier challenges so if basically if we can get the chips we can do it so hopefully this chip shortage will alleviate soon but i feel confident of being able to maintain something like this at least above 50 for quite a while the model 3 became the best-selling premium vehicle globally so if any premium vehicle

1:41but i i mean i almost got arrested at one point for claiming that we'd do 5000 a week literally laughing now anyway this is great we think the model y will be the best-selling vehicle of any kind globally so i think it will exceed the model 3. i think we've got a good chance of it being the best-selling vehicle by revenue next year and then i think quite likely to be the best-selling vehicle in just of any kind numerically in in 2003 and 2023. so basically we need we need awesome to get online and berlin to get online and reach volume production and then i think that's going to happen in terms of a free cash flow generation obviously we had some tough years

2:42back then things were looking a little dicey to say the least in 2017 and 2018 don't want to go back there again but we got through that and now things are looking really good so i think we'll see continued strong cash flow generation and especially as you multiply unit volume times autonomy and increased efficiency in the factories because i think over time you'll see all manufacturers will make electric vehicles and eventually all manufacturers will make autonomous vehicles and we i think and tesla's open to licensing autonomy because i think autonomy will be such a significant lifesaver and preventer of injuries that it is not a technology we want to keep to ourselves

3:43so i think it will be morally right to to license it to other manufacturers if they would like to to use it so [Applause] and of course we made a lot of progress on on cost reduction and so despite our average selling price actually going down significantly because with the introduction of the model 3 and model y these were much lower priced cars we've managed to still do decently well on on gross margin so this is you know getting the average price down and and gross margin up is very difficult but we've managed to to do that so yeah it's good our goal really is to make the cars as affordable as possible we are seeing significant cost pressure in our supply

4:42chain and so we've had to increase vehicle prices at least temporarily but we do hope to actually reduce the prices over time and make them more affordable so yeah unfortunately we've just expedited like i mean the sheer amount of money we're spending on flying parts around the world is is just not not great but hopefully temporary so and we need a lot of batteries hence a battery day that's what this shirt means very obtuse but we are gonna need a lot of batteries this is gonna be a combination of batteries from our suppliers and in supplier discussions you know some of our suppliers have just asked me outright are we going to just you know put them out of business or

5:31something i'm like not at all as many sales as you want to make and supply to us at an affordable price we will buy no limit i'm like oh okay so like did you want to have yeah increased by 100 sounds good to me so the basic plan is we're really going to order a lot of we have ordered a lot of batteries from suppliers basically telling suppliers literally go go as much as you can make will take and and we'll we'll prioritize batteries for vehicles but then use any excess cells that we have in the power wall and mega pack because over time we think the demand for stationary storage is going to be at least as high as the demand for vehicles so sustainable energy primarily solar and

6:24wind is intermittent and so the wind doesn't blow all the time the sun doesn't shine all the time obviously and so you need batteries to buffer that power so the fundamental pillars of a sustainable energy future are basically solar and wind those are the primary stationary batteries and electric transport and if you have those three then you have a sustainable energy future as long as the sun is shining you know so sometimes people ask me about fusion and i like it as an idea and by the way i think it's totally doable but there's a giant fusion reactor in the sky that shows up every morning and zero maintenance so i'm like sounds like a good deal you know

7:05or just catch a little bit at sunlight and power earth a shockingly small amount of land is needed to power earth it's like you know a couple hundred miles by a couple hundred miles of solar panels will power the entire united states so it's like okay it's not that hard and then i believe we calculated you only needed one square mile of batteries so you may think like wow a lot of batteries one square mile of batteries it's not that crazy so anyway so we've got we've got a plan to reduce the cost per kilowatt hour of batteries and and our suppliers you know have similar plans so this is really supplemental to our suppliers we'll make cells they'll make sells

7:59we'll use them all the the fundamental good of tesla i think is by how many years did we accelerate sustainable energy this is the fundamental i think way to think of the the value of tesla and so if we are able to accelerate sustainable energy by more years that is good hence the need to grow quickly we've got three new factories jaeger shanghai has done an incredible job and changing our now exceeds fremont in production so actually i'd like to just give a special hand to the tesla china team [Applause]

8:50so it's the best quality lowest cost and and also low drama so it's great and that said we are continuing to expand our fremont operations and expect to hopefully increase fremont output by 50 so and that's still where we make all model s and x's are made in fremont but it kind of makes sense especially for the high volume vehicles to have production that's at least on the continent where the consumers are otherwise it's just it's also not good for the environmentally shipping cars you know several thousand miles so the basic idea is have the high volume vehicles be where the customers are approximately at least and then also great progress with building giga

9:42texas which is where we are right now and giga berlin brandenburg so just hand for those teams as well [Applause] and and the these factories will have cell production in them as well so this will be really kind of raw materials in cars out so really really big yeah i mean these things will be in like units of pentagon basically

10:19let's see so impact report there's a lot of i think interesting stuff in our impact report we go into i think quite a lot of detail on all things we're doing and you know a tesla is is certainly a company that tries very hard to do the right thing in all respects we try very hard to do the right thing in all respects we don't always succeed but if you're looking for a company where you say is that company really trying to do the right thing that is tesla okay we really try

10:58okay i mean that's there anyway so so as i was mentioning we've got the you know three parts of a sustainable energy future solar and wind but i think primarily solar will be the the main source of energy of sustainable energy and energy in general and then you need to store that energy with a stationary battery packs and then you need electric vehicles and electric airplanes and boats and whatnot so yeah great

11:42yeah and then the average life cycle missions in the u.s this is only going to get better as as we move to a sustainable energy grid and electric vehicles then obviously we move to a fully sustainable energy economy which is where we want to get to as quickly as possible the sooner the better and can there be a carbon tax i mean what the hell you know so people someone said like or carbon tax that would benefit tesla i'm like yeah but it would it would hurt spacex so how about the carbon tax it's really needed so see battery materials are definitely incyclable button gas is not co2 is an extremely stable molecule mars atmosphere has been primarily co2

12:32for billions of years it's extremely stable so sometimes people worry about methane do not worry too much about methane methane quickly breaks down into co2 methane is not a stable molecule co2 is extremely stable so and of course you can recycle battery materials so you can think of batteries as essentially high-grade ore so you can either get your lithium and your nickel and the various constituents or the battery from rocks or from batteries it's much better to get them from batteries so batteries are basically high grade ore and tesla's already started recycling and their electric companies are going to do recycling because it basically pays to

13:18do recycling for batteries so we're seeing you know increased extreme weather events and there's like wildfires and here in austin there was a massive snowstorm that turned the power off i was actually in austin for that snowstorm in a house with no electric no lights no power no heating no internet couldn't actually even get to a food store if you could get to a food store there's no food there that's went on for several days however if we had the solar plus power wall the car would have had lights and electricity and actually had a starlink internet you'd have internet too so

14:08all the things you need for prepper basically doomsday comes yeah it could be helpful so in factory safety we've done we've made huge improvements on on factory safety so we're now 18 percent better than the industry average so this is a this is great it's always tough with safety as you ran for production lines and as you start up factories but then as as once it's in steady state then the injury is naturally naturally declined because people get used to it and you iron out the issues and so we're seeing excellent factory safety in tesla and our goal is to have the safest factory on earth and then ai day i think it was important to change the fundamental perception of of

14:59tesla because people do they sort of think of tesla as a car company and yes we made cars but but but the ai portion part of tesla was not well understood i tesla is as much a software company as it is a hardware company and and and we also do the the chips so we designed the the full self driving inference computer we're designing a training computer that's going to be able to we think be the most efficient neural net training computer in the world by far and and we're seeing a tremendous response so daily applicants by roll as you can see it's basically you know on the y-axis there and then after ai day that's the ai applicants increase dramatically so i

15:54thought that was a very successful day team did a great job and yeah ai is going to be a very important part of the future self-driving is obviously one of the functions and i also have mixed thoughts about ai and we've got to watch out for ai being a danger but it's happening either way so i guess if we help do it we can try our best to make sure it is a positive you know good good ai hopefully yeah and then i'm excited to announce that we're moving our headquarters to austin texas [Applause]

17:01so just to be clear though we will be continuing to expand our activities in california so this is not a matter of of sort of tesla leaving california as i said we're our intention is to actually increase output from fremont and from kicking nevada by 50 percent so but we're just we're just hitting the sides of the bowl you know i mean if you go to our fremont factory it is jammed i mean it jam it's like whoa and

17:32when we first went in there it was it was like you know well we're like a kid in his parents shoes it was ridiculous like tiny s and giant factory and now we're like spamming a can here like how do we put more stuff and and it's it it's it's tough for people to afford houses and a lot of people have to come in from far away and so it's we'll take you're taking it as far as as possible but it's there's a limit to how how big you can scale in in the bay area so here in austin and our you know factories like five minutes from the airport 15 minutes from downtown and we're going to create an ecological paradise here on the car because we're around on the collar of

18:14colorado river it's going to be great so temperature size gain continuing to expand in california significantly but but even more so here in texas so

资源与出处

相关视频

同频道、同主题或相邻年份