特斯拉最大的长期竞争优势是什么?
2018 年 6 月的年度股东大会上,马斯克给出的答案让台下一部分人愣了一下:制造。
一家靠设计、软件和品牌起家的公司,说自己最强的护城河在工厂车间。更反常识的还在后面,他紧接着承认,特斯拉在 Model 3 上把这件事搞砸了。
这场一个多小时的问答信息密度很高。Model Y、Semi、Roadster 的时间表,电池成本的临界点,上海工厂,自动驾驶的路线之争,3.5 万美元版本的承诺,还有他对“马斯克时间”的坦白。下面按线索捋一遍。
我们犯过一个特别蠢的错误
Model 3 产能爬坡最艰难的那段时间,工厂里出现过一种画面:一些原本人两秒钟就能做完的动作,被交给机械臂,机械臂笨拙地折腾半天。
马斯克的原话是,看到那个场景会觉得特别蠢,为什么我们要这么干。
问题出在过度自动化。他承认特斯拉在 Model 3 上对自动化程度的判断过于乐观,还大量依赖一级自动化集成商,其中好几个项目彻底没跑通。
教训很清楚:有些工序特别适合人做,有些特别适合机器人做。把适合人的活交给机器人,是最典型的错误。
正确的顺序反过来。先建一条偏向人力的产线,让它跑起来,然后把那些对人最痛苦、最枯燥、最容易造成重复性劳损的环节逐步自动化。焊接、搬运、喷涂这类活交给机器,线束穿插、内饰装配这类需要手感和判断的活留给人。
这套思路会用在此后的所有工厂上。他还宣布,特斯拉要把一级制造系统全部内部化,供应商退到二级、三级。这个决定听着不起眼,实际影响很大,意味着特斯拉要从“集成别人的方案”转向“自己定义生产线”。
把制造当成核心竞争力,这话从一个以产品经理形象示人的 CEO 嘴里说出来,分量不太一样。车可以设计得很漂亮,软件可以 OTA 升级,但最后能造多少辆、多便宜、多快,决定公司能不能活。2018 年那个时间点,特斯拉正卡在这道关上。
电池 100 美元这道坎
电池成本是电动车行业最敏感的数字,特斯拉一向不公开。这次 JB Straubel(特斯拉联合创始人、时任 CTO)在场,也只肯给区间和方向。
他的说法是,特斯拉仍然确信自己的电池在价格和性能上是全球最好的。如果外面有更好的,他们不知道。为了确认这一点,团队找过每一家电池初创公司、每一个实验室、每一家大型制造商,拿报价,测电芯。这种地毯式的排查,本身就是一种态度。
然后是两个具体数字。
电芯层面,低于每千瓦时 100 美元,可能在 2018 年内就能做到,前提是大宗商品价格不大幅波动。
电池包层面,也就是整车真正关心的那个数字,长期能做到低于每千瓦时 100 美元。马斯克补充,这里的“长期”指不超过两年。
路径是三件事叠加:电芯化学体系继续改进,生产工艺继续改进,以及把正极和负极材料的生产也搬进超级工厂,做更多垂直整合。再加上模组设计的优化。
100 美元每千瓦时是个心理关口。跨过去,电动车和燃油车的成本对比会发生质变。当时行业内不少机构的预测是 2025 年前后,特斯拉给的时间表整整提前了几年。
Straubel 还提到一句,他们在能量密度和电池包成本上已经有了几个“相当酷的突破”,只是不便细说。这种话在发布会上听听就好,但由他讲出来,可信度不低。
上海工厂,和未来的 10 到 12 座
这场大会上,特斯拉第一次公开确认中国工厂的选址:上海。
全球销售负责人 Robin Ren 上台宣布这个消息,措辞是“我们有非常好的讨论,非常好的合作伙伴”。具体细节留到下个月。
马斯克的补充更直白:要让车变得让人买得起,生产本地化至少要达到洲一级。北美一座大的整车加电池工厂,中国一座,欧洲一座,这是显而易见的三处。理由不复杂,车重、体积大,跨国运输成本高,关税又是一层,本地化生产是规模化的前提。
中国工厂的细节下个月公布,欧洲工厂大概今年年底定址。他判断,最终全球会有 10 到 12 座超级工厂。
内华达那座现有的超级工厂,还要再建四五年。它已经是世界上最大的建筑之一,按规划最终可能是第二名建筑的两倍大。目前完成度大约三分之一。
未来的超级工厂形态会变:整车、电池包、动力总成做成一个集成单元,不再分开建。供应链的几个环节因此压进同一栋楼,运输、库存和协调成本都能省下来。
Semi 和 Roadster 会比发布时更好
关于 Semi 电动卡车,马斯克说会再做一次设计修订。理由是在过去这段时间里学到了很多,现在有把握把它做得比发布时更好,续航会远超行业里大多数人认为的可能。
目标市场也讲清楚了:这款车要在欧洲、北美、中国和世界其他地方都能用。这意味着充电标准、法规认证、道路条件都要一并考虑,产品从设计之初就面向全球长途干线,美国只是其中一站。
Roadster 那边有个更具体的彩蛋。发布会上展示的性能,只是基础版。会有一个 SpaceX 选装套件。现场听到这句的时候掌声很大,说明这个说法戳中了某些东西。
为什么要把 Roadster 做到那个程度,他给了一个跟销量无关的理由:燃油车有一种光环效应,好像天生就比电动车高级。如果能在每一个维度上都超过燃油车,这个光环就会消失。对普通公众的认知来说,这件事很有力量。
这番话把 Roadster 的定位讲透了。卖车只是表层,真正的目标是改写整个品类在人心里的位置。一款车卖几千辆,改变的是公众对电动车的默认认知。
产品线的时间表大致是:Model Y 大约在 2019 年 3 月亮相,之后大约两年进入生产。Semi 和 Roadster 的节奏类似。这段表述在文字稿里有些含混,语音识别可能吞掉了几个词,但“先亮相、后量产、中间隔约两年”这个结构是清楚的。
3.5 万美元的 Model 3,说到做到
3.5 万美元版本是 Model 3 发布时最核心的承诺,也是被追问最多的一项。
马斯克的回答很干脆:一定会提供。时间上,今年年底应该能做出更小的电池包,然后 2019 年第一季度进入 3.5 万美元版本的量产。
他补了一句:如果物理上可行,现在就做了。这句算是承认了延迟,也把原因指向了产能本身。先把高配版本的产能拉起来,才有余力做低配。
另一个被反复提到的数字是保险费。特斯拉正在和保险公司合作,也有内部动作,目标是把 Model 3 的保险成本做到比宝马 3 系或者同级别中型轿车低至少 20% 到 30%。安全性更好是一方面,维修成本也在往下压。
总持有成本,会是 Model 3 相对同价位燃油轿车的一张牌。购车价只是其中一项,保险、保养、能耗、维修加起来才是真实开销。
产能方面,他给了一个很具体的画面:总装环节目前有两条线,第三条两周前开工,已经有第一台车下线。这条线的设计比前两条好得多。
现有的两条线大概能撑到每周 5000 辆,加上第三条,他有信心超过 5000 辆每周。这个数字是当时市场最关心的指标,因为它直接决定季度交付和现金流。
等待时间分市场。美国现在下单,大概三到四个月。右舵版本要一年以上,因为要建产线、运过去、还要逐国做认证。整体区间三到十五个月。
试驾车本月底开始铺,下个月底覆盖美国大部分门店。服务中心一年内翻倍。钣喷是另一个痛点,特斯拉自己在建钣喷中心,本月底覆盖美国前十大都会区,目标是很多维修当天完成。第三方最快也要一周,有时几周。
这个细节容易被忽略,但对车主体验影响巨大。车卖出去只是开始,维修网络的密度决定口碑。
租赁会在今年年底或明年初上线。晚一点的原因很实在:租赁会占用特斯拉的资本。当时公司的现金状况经不起这种占用。
关于激光雷达,马斯克的一句话很重
有股东问,自动驾驶快到终局了,用激光雷达和不用激光雷达的路线迟早要见分晓,怎么看。
马斯克的回答是,激光雷达会把公司引向一个局部最优。它像拐杖,能帮你走到接近终点的地方,但只要你依赖它,就永远到不了终点。
要让视觉做到极致,才能真正实现自动驾驶。视觉做到极致之后,激光雷达就不必要了,它不增加任何东西。
他还讲了技术细节。车上有超声波传感器负责近场,有前向雷达负责雾、雪、雨这类低能见度场景。而激光雷达在这些场景里恰恰是失效的。
更根本的质疑在波长。激光雷达是可见光波段的主动光子发生器,波长在 400 到 700 纳米。可环境里本来就有海量可见光子,在这个波段主动打光,信噪比很差。他认为应该瞄准 4 毫米左右的波长,那个波段能穿透遮挡。
这套判断对不对,行业里争论很大。但它解释了特斯拉为什么坚持纯视觉加雷达的路线,而且愿意为此承担风险。多年之后回看,这段论述是整个路线的原始表述之一。
自动驾驶的进度,他给了两个时间点。高速匝道到匝道的自动通行,大概两个月内能推出来。他前一晚凌晨一点还在测,用的是简化版控制系统。
过去团队同时走两条路,一条很复杂,效果不好;一条简单,看起来能成。这种“承认自己走错路”的表述在这场问答里出现了不止一次,是整场讲话的一个底色。
本周会推一个新版本,他认为提升明显。接下来六到十二个月,autopilot 的可靠性和能力会快速上升。下个月还会开放免费试用,让车主自己体验。
还有一个细节来自现场一个小女孩的提问。她说妈妈开 Model 3 靠近前车或停止标志时,车会大声报警,能不能有个模式让车提前、平缓地减速,而不是最后一刻急刹。
马斯克的回答是,会有。系统会按照置信度比例调整速度,置信度高就多减速,置信度低就少减速。难点在于不能烦人,很多时候车以为要减速,其实不用,那会把驾驶员逼疯。这需要神经网络和配套规则做得足够精细。这个改进会随今年晚些时候的版本落地。
这个问答是全场最生动的一段。一个孩子提的问题,恰好戳中了自动驾驶产品化的真正难点:技术能做到,和让人愿意用,是两件事。
充电、续航,以及一个披萨大小的终端
超级充电的下一步是 V3。马斯克提到,V3 会更多采用太阳能加电池的集成方案,长期目标是几乎所有超充站都配太阳能和电池。
好处是超充站可以脱离电网独立运行。有太阳、有电池,哪怕电网断了也能充电。他用了一个开玩笑的说法,就算僵尸末日来了,只要太阳还在,超充站就能工作。
充电速度能提升多少?数量级谈不上,三到四倍是可能的。而且这个提升不适用于所有电池化学体系,2012 年的电芯扛不住现在的充电功率。
他给了一个挺有意思的判断标准:行驶时间和充电时间的比例。如果能做到 8:1 到 10:1,也就是开十小时充一小时,那人类的生理需求反而成了瓶颈。
现实一点看,一次中途休息大概半小时,这是让车在你回来之前准备好的最低门槛。如果充电时间能压到 15 到 20 分钟,车会比人先准备好。这个思路把“充电慢”重新定义了一遍:参照系换成了人的休息节奏,加油枪的速度不再是唯一标尺。
续航方面,他希望未来提供超过 400 英里的车型。Model S 100D 小心开的话,已经能从洛杉矶一口气开到旧金山。
关于 Starlink,有人问特斯拉能不能用卫星网络。马斯克的回答比较务实:Starlink 主要面向固定的家庭和企业,终端大概是一个中小号披萨那么大,装在车顶不太合适。车还是继续用 Wi-Fi 和蜂窝网络。
虚拟电厂是另一个话题。澳大利亚有个项目,把最多 5 万户装了 Powerwall 的家庭联网。每家有电池,停电时能当备用电源;正常时这些家庭可以互相通信,再和电网沟通,整体表现得像一座分布式电厂。
美国有小规模试点,之后会往全球推。波多黎各那边,特斯拉建的电池和太阳能项目比所有其他公司加起来还多。当地电网在飓风后几乎瘫痪,分布式能源的意义在那样的场景里格外具体。
超充站加太阳能的思路,和虚拟电厂是一套逻辑:把能源的生产、存储、消费尽量放在本地,减少对集中式基础设施的依赖。
两个数字:30% 和翻倍
电池能量密度还能提升多少,是所有人关心的问题。
马斯克的回答分了两层。同样大小的电池包,能量提升 30% 是能看到的路径,运气好能到 40%。这个时间尺度是两到三年,指的是体积能量密度。
翻倍需要六到八年。
真正的分水岭在锂金属负极。用纯锂做负极,是那个大跳跃的关键。这条路的技术可行性他有信心,难的是规模化和可靠性。实验室里做出来和工厂里稳定生产,中间隔着一条很长的路。
关于回收,他的态度很明确:特斯拉会回收所有用过的电芯、模组和电池包。现在在每个有车辆运营的大洲都有合作公司,同时内部也在做工艺研发,目标是提高活性材料的回收率。
终点是闭环,超级工厂直接用回收材料。这条路能看到,现在还没走到。长期看,回收对成本的好处很大,能少买很多新原料。钴、镍、锂这些材料的开采和价格波动,一直是电动车成本里最不稳定的部分。
顺便,现场有动物保护组织的代表提问皮革使用。马斯克的回应是,Model S、X、3 目前只有方向盘轮缘用皮革,需要全素车的客户可以通过“菜单外”的方式更换。Model Y 会完全不用皮革,包括方向盘。
这个回答没什么戏剧性,但它展示了一种处理质疑的方式:承认现状,给出过渡方案,说明下一步。
马斯克时间
这场问答里最有意思的一段,来自一位长期投资者。他说自己不愿意这么说,但 Model 3 的产能爬坡让他对特斯拉的时间表信任度下降,以后是不是该继续按“马斯克时间”打折。
马斯克的回答没有辩解。
我在时间上确实有点问题。我兄弟在这儿,小时候我们一起赶校车,他会骗我时间,总说得比实际早一点,结果我们反而总能赶上。(马斯克)
他接着说,自己天生是个乐观的人,如果不乐观,大概根本不会去造车和造火箭。他正在努力重新校准这些估算,也许应该给未来的日期留一点余量。
这段话值得单独拎出来。那些总是跳票的时间表,背后是一种结构性的乐观。这种乐观既是特斯拉能做成很多事的原因,也让它一次次让市场失望。
投资者该把它当成一个需要建模的变量,普通用户可以拿它调整预期。听到“两年后量产”,心里换算成三年半,误差就小很多。
还有一些零散的问答,信息量不大但有意思。
有股东问会不会做摩托车。马斯克说小时候骑越野摩托骑了八年,17 岁前有辆公路车,差点被卡车撞死,所以不做。这个回答很马斯克,用个人经历直接否掉一个商业问题。
有人问紧凑型车。答案是五年内。这是他第一次给出这个品类的明确时间窗口。
有人问 JB Straubel 为什么从第一季度的风险因素里消失了。马斯克说这是疏忽,本意是让措辞更通用,没有任何暗示。他当场说了句“JB 我们爱你”。这种当场澄清,对一个高管的去留传闻来说,是最直接的处理方式。
有人问 2012 款 Model S 和现在的差别。马斯克说 Model S 现在大概已经是第三或第四版,改进幅度巨大,建议早期车主去试驾现款。他承认特斯拉在“让老用户知道车进步了多少”这件事上做得不好。
还有储能业务的毛利率。当时还是负的两位数,他的目标是做到和汽车业务同样的水平,20% 到 30%,时间点是今年晚些时候或者明年初。他拿 Model 3 和 Model S、X 的早期阶段做类比,说负毛利是产能爬坡的必经阶段。
留下能用的几条线索
把整场内容串起来看,有几条线索值得记住。
第一条是制造。特斯拉花了很大代价才明白,自动化不是越多越好,顺序比程度更重要。这个教训对所有做硬件的人都有用。
第二条是成本曲线的位置。电芯 100 美元、电池包 100 美元、能量密度提升 30%,这三个数字构成了当时特斯拉对未来的基本判断。它们能不能兑现,直接决定电动车什么时候真正平价。
第三条是路线选择。纯视觉加雷达对阵激光雷达,是一场持续多年的争论。马斯克在这场问答里给出了他最完整的论证,包括那个关于波长的技术细节。
第四条是预期管理。一个承认自己有“时间问题”的 CEO,给出的每一个日期都该打折听。跟道德无关,纯属使用说明。
今天读这份文字稿,承诺兑现没兑现已经不那么重要。更有意思的是看一个人怎么在公开场合承认错误、解释判断、给出路径,然后再一次给出过于乐观的时间表。
- 如果你在评估一家硬件公司的路线,先看它对制造的理解,再看它的产品设计。前者决定能不能规模化,后者决定有没有人买。
- 听到任何“两年后量产”的说法,先问三个问题:产线在哪、良率多少、供应链是否已经锁定。三个都答不上来,时间表基本不成立。
- 看技术路线之争,别只看结论,去看双方对失败模式的分析。马斯克反对激光雷达的论证之所以值得读,是因为他具体说了激光雷达在哪些场景下失效,而没有停留在“不好”两个字上。
- 判断一个 CEO 的预测可信度,把他过去三年的公开时间表和实际交付时间列出来,算一个平均偏差系数,以后直接用这个系数调整他的新预测。
- 买车或者买任何耐用消费品,把总持有成本算进去:购车价、保险、保养、能耗、维修、残值。发布会上讲的通常只是第一项。
0:00model why approximately March next year and then go into production about to dig around two years from now maybe a little less than two years but basically first half of 2024 production model while something similar for semi and and roadster so these these products are shaping up I think they were semi and roadster are actually gonna be even better than what was unveiled we've figured out ways to to improve the the range and overall functionality of the semi in particular the Roadster we what
0:40I unveiled with the Roadster was the base model performance that's it's it's going to have a SpaceX option package [Applause] it's crazy and I think it's important for us to show with the Roadster that an electric vehicle can outperform an gasoline car in in every way so that because gas in coastal have sort of a halo effect and I think if we can show the electric car can outperform gas in car in every way then we sort of get rid of that halo effect of gasoline cars and and I think that's quite a powerful thing perceptually for the for the general public [Applause]
1:42so these are some questions that we got on Twitter I guess it's a really insightful questions on on Twitter as well as some strange ones but yeah so with most questions is in in from wealthy production line as I said before we're really going to focus on manufacturing technology for Tesla and we've made a lot of mistakes with the model 3 production that but we've we recognize those mistakes and we're confident we know how to address them and we factory what we are addressing them and at and long-term I think the
2:24the biggest competitive strength of Tesla is going to be manufacturing this is this is sort of counterintuitive but it is it's gonna be quite dramatic I think the the approach to automation that we're taking I think it's some some cases has worked in some cases not but it's it's clear that that there are there some elements of production which are really well suited to people doing it and some parts production that are really well suited to robotics and you
3:03know one of the biggest mistakes we made was trying to automate things that are super easy for a person to do but super hard for a robot to do and and when you see it it looks super dumb Wow why do we do that so and then as you sew it sort of makes sense to have start off with an initial production line which is has a relative bias towards towards
3:29people and then you automate the parts of the production system that are the most painful and difficult for people to do so it particularly wants that result in repetitive stress injuries or mechanically difficult but that's that's really a much better approach and that's that's what we're gonna do for general factories wanted to it's much more sensible way to do things see see well let me actually have foot
3:57Tesla so much has executive team come up guys you want to come up and hang out [Applause]
4:22JB joint Oh about the battery Kostov I'm sure so you know it's difficult for us to talk about specific cost numbers that's always a difficult topic but we are we're still very confident that we have the best price and performance of anything out out there in the world if there's something better I don't know about it and we've looked as hard as we possibly can and we try and talk to every single battery you know start up every lab every large manufacturer we get quotes from them we test cells from them so if there's something better you know we're all ears we'd love to find it but we haven't found it yet so generally yeah we're still pretty confident about that same direction yeah
5:05we think at the cell level probably we can do better than $100 per kilowatt hour maybe later this year depending upon what on commodity prices if commodity prices are roughly where they are today then we'll probably do better than $100 kilowatt hour at the cell level and then with further improvements to the cell chemistry and the production process and more vertical integration on the sell side for example integrating the production of of cathode and anode materials at the gigafactory and then an approved design of the module impact we we think long term we can we can get below $100 a kilowatt hour at the pack level which is really the key figure of Merit for a a car but
5:57long term meaning definitely less than two years that has a long term anything I mean yeah yeah we think we think we've caught with some pretty cool breakthroughs on this front but on the energy density and end cost of the battery pack and yeah I think it's going to be pretty pretty great shorty
6:30when will the gigafactory be completely built I think we'll keep building on the gigafactory for at least four or five years it will be by far the biggest building in the world it's not that far from being the biggest building in the world already and based on the plans do we know it
6:55might be as it might be twice as big as the next bullying world it's the interesting tour so it's yeah yeah it's about a third down right now and yeah so it's really really really enormous and and I think it's going quite well there will be more Giga factories in the future like we're close to announcing a
7:25combined vehicle and battery factory so future gagger factories will include vehicle and battery pack and powertrain as a single integrated unit and we're close to announcing something in China that I don't want to talk about that but we don't make an announcement exactly it may be just talked about like the preamble or something so Robin is
7:59the head of worldwide sales Potenza thanks to you long I didn't expect to to talk about this so we are incredibly excited to to build first Tesla Giga factory outside of the US in China is specifically it's going to be in Shanghai and we have been holding discussions with the government various governments in China really great discussions great partners we really look forward to to to working with them in the years to come this is going to be you know the the next generation of Tesla factory who is super excited the stuff that we're going to be put in there and the cars that we're going to be building in that factory is going to be incredible
8:54so we get a announced something really all the details really really soon so I won't tell more but this is enough thank you it particularly has returned to make cars more more affordable it's getting important to localize production to at least the continent level and so having a big factory of vehicle Factory in North America one in China and then one in Europe will be necessarily up is the obvious three places for vehicle and
9:35battery Giga factories so probably things go according to plan we'll be announcing the details of the the China keg factory as soon as next month and then Europe Giga factory maybe end of this year it's kind of depending on on how the we need to figure where to put it exactly so but probably towards the end this year for the your fkk factory and altom utley we expect probably there's 10 or 12 worldwide so if some
10:16questions on the Tesla semi we we are going to do another revision of the Tesla semi design because this we've learnt a lot and we think we can actually make it even better than what was unveiled and and really have a range that is way beyond what people think what most people in the industry think is possible and we want to it's definitely gonna be a semi that works in Europe and in North America and China and the rest the world here was
10:49convinced that that yeah one major Factory in the works and then another one in the works later this year yeah so what sells yeah we've we've talked about this a few times but Tesla will absolutely recycle and we do recycle all of our spent cells modules and battery packs so you know the discussion about you know this waste is sort of ending up in landfills is not correct you know we we would not do that these are valuable materials in addition to it's just the right thing to do so so we have current they're companies on every major continent where we have you know cars operating that we work with to do this today and in addition we're developing
11:32internally more you know processes we're doing R&D on how we can you know improve this recycling process to get more of the active materials back and ultimately what we want as a closed-loop right at the gigafactory x' that reuses the same recycled materials this isn't impossible we see a pathway to do it but you know that that's where we're headed with this and you know today the you know we're on the way to do that it's definitely a something that'll be a huge benefit in the long term to cost as we're able to reprocess more materials instead of actually having to buy new materials
12:15it says repairability of multi including insurance costs we're working with the with insurance companies and on some internal activities they were really confident of getting the ocular cost of insurance for model three to be at least twenty to thirty percent lower than say a BMW 3-series or equivalent mid-sized sedans so the the safety is definitely better and then we're we're working on the the repair costs we were made a lot of progress in that front but the bottom line is that the the the insurance cost will be total cost cost ownership of them all three should be significantly better than any other mid-priced premium sedan so yes there will be
13:10we will definitely offer a thirty five thousand dollar version of the model three and it's I think probably at the end of this year is where we should be able to make the smaller battery pack and and then get into kind of volume production of 35k version in q1 next year so that's definitely we will definitely honor that obligation and we would do so right now if it was physically possible see yeah we're gonna
13:44in in hopefully next month offer a free trial for people to try out autopilot and see how well it works we're also making rabbit progress rapid progress on autopilot technology so the there's a new version of autopilot that's rolling out I think this week which i think is quite a significant improvement and I think the what what you'll see is that the reliability and capability of
14:15autopilot will increase exponentially over the next six to 12 months it's really the improvements are very very rapid the length of time to wait for a model three if you if you're ordering one now will vary quite a bit depending upon what part of the world you in you're in and what configuration so if for the existing configuration if you were to order now in the US you probably would be waiting I'm guessing about three or four months if you thought only other had if you want the right hand drive version you're probably waiting for over a year because we need to build the right hand drive version and ship it to other countries and homologate the car
14:59for four other countries so the weight is anywhere from three to 15 months approximately but for current configuration ordered now it's made about three or four months this is actually quite a complicated answer there are many lines to the model model three and some faces it's there are several lines in some places that are there's just one line and it kinda depends on what the cable the capacity of that line is so for for general assembly which is like putting the parts together at the end we currently have two lines and are constructing a third the the third line is I think
15:45dramatically better than lines one and two we started construction on on that third line about two weeks ago and we're already putting the first car through that line so it's really crazy fast and that's part of what gives me confidence about the 5k per week for model three but currently the biggest constraint on output is general assembly and I probably get to five thousand a week with the current to general assembly lines but with the third one I'm highly confident that we can exceed five thousand units per week oh and then
16:27model three test drives we should build off of all three test drives starting at the end of this month and I think we should have them in almost all stores you know America by the end of next month we're rapidly expanding service centers I think by the end of the kind of year over a year probably see a doubling of service center capacity for Tesla and
16:58we're making major progress on the body shop front we're all set it was quite a quite a big deal we're creating Tesla body shop repair locations and we should have by the end of this month of the at least the top ten metro areas in the u.s. being able to be serviced by a Tesla body shop that this will be a dramatic improvement in the cost and time of flood repair in fact we think we might be able to do for a lot of them like the same day body repair which is yeah it's definitely
17:42possible so I think I think we want to aim for it'd be some number of bloody repairs to be the same day whereas if we go to the third parties best-case it's about a week and some cases at several weeks so this is this is pretty exciting actually and we're basically just taking out biggest service centers adding in an annex full body repair and and then pre
18:06stalking the parts so you don't have to wait for parts to come from the factory not yet but yeah March next year I think we'll have something very exciting show all right so let's take some questions from the audience I think it's basically just line up at the mics and go for it
18:38is there gonna be a time you know perhaps in the next and Tesla is going to produce compact and or subcompact vehicles such a huge segment and it seems like that would be necessary to fulfill the master plan part do mandate yeah I think we'll do a compact car in less than five years yeah yeah great
19:05progress you mentioned out of pilot progress as well so when do you expect like fully enhanced autopilot functionality like following navigation switching freeways within the six-month timeframe I was just testing that last night at about 1:00 a.m. I think we might be able to release something in a couple of months that can do that we've
19:38been pursuing two paths one really complicated path that I think isn't working that great and then a simple path that I think will work pretty well I mean I was able to drive last night go from highway on-ramp to highway off-ramp using the simplified version of the control system and I think with some with them further effort we can get that out in the next couple months yeah this
20:12on there we go as far as testlink is concerned with the model 3 and to what degree of certainty will there be a consumer or a fleet lease option within the next three years on the model 3 is that something that you guys already have in the plans and works or is there room for ancillary business we do we will offer leasing on
20:38model 3 but probably end of this year or early next because it does have a slight impact on the capital usage of Tesla we you know it tons of fleet stuff I think we yeah I think people can certainly buy a lot of model threes and and then operate them as a fleet like people do for Model S's and X's for taxis yeah I
21:10think we're so maybe happy to support that simple follow-up to that just kind of the understood thing if I purchase a whole fleet and then Tesla comes in and says all right we're gonna start leasing direct to the consumer or having a commercial option that might not work out so well for me but that that's why I'm asking is it anything commercial
21:39happening that you have in the works in the next two years even well we don't really think right now we're just super focused on ramping up manufacturing of model 3 and make sure people can get their cars because they've been waiting and waiting for a couple years and we're not really thinking much about incremental incremental demand generation because as it is even getting
22:06to 5k cars per week we would it would take us almost two years to produce enough cars to satisfy those that are put down $1,000 deposit so yeah we need to kind of ramp to 5k and then next year ramp to 10k a week and get the right hand drive version done and I'm Allah gate the car for Europe and Asia and you know and then we'll
22:34think about other things once you've done all those things thank you all right Gwynne Shotwell if spacex mentioned that tesla two mobiles might use in some way the StarLink satellite network I was wondering if you might elaborate on that opportunity and when that might take place it's possible I think it probably
23:02will be the startling thing is more meant for like like fixed terrestrial homes and businesses and that kind of thing for mobile it might be possible to use the Salling system effectively if you had a repeater ground-based repeater system but the styling used terminal is about the size of a sort of small to medium sized pizza so I'm not sure you'd want to put that on the roof of the you know for Tesla might maybe but
23:40I think probably just using what's like we will continue use just Wi-Fi in the cellular network yeah what's like we hello all right my name is Steven singleton and my question is how is Tesla engaging regarding virtual power stations with governments and territories in countries that may have weak power infrastructures to provide clean energy to more of the world citizens yeah I think the will have a
24:17lot more to say about that when we announced the generation 3 of the superchargers because that will be doing much more of an integrated solar battery system with the superchargers so today only a few of the supercharger systems have solar and battery systems but long term we want to have almost all of them have that and the nice thing is that if you've got like a solar powered kind of like a solar powered carport area and and in Tesla batteries then even if the crew is you don't even need to be connected to the grid so it's sort of like even you know proof against like a zombie apocalypse so it should still work as long as there's always on tuned
24:56year the supercharger I suppose but yeah desert it in it'll be able to work anywhere even if there's not good power infrastructure and for like a virtual power plant if that's kind of what you also were alluding to you know we had we do have a really cool project in Australia where we're actually networking together up to 50,000 individual homes with power walls so each one of those homes has its own battery it can still serve as like a you know backup power source if the utility totally goes out if there's a storm but when things are working normally all those houses can talk together and then we can talk to the utility and treat them as sort of one big distributed
25:38power plant so that that's a really cool project that has benefits across the whole grid for the homeowners for a lot of people and you know we're building that out right now and will probably be expanding that same model we have a small demo in the US and we'll be expanding at world
25:59[Applause] and we're also doing quite a lot in Puerto Rico I think we have I think we tells us more put more battery projects and solar projects in Puerto Rico than everyone else combined so it's making a big difference they are doing our best to and I think this potential for kind of a virtual credit in Puerto Rico as well rather than rebuilding a legacy you know some oil and gas based energy generation system
26:33Halon I also think boring boneheaded questions are not cool so hopefully this is a little more interesting I've added Model S since 2012 best purchase ever thank you guys nail design nail performance the one thing I always get from friends and family because I do the drive down to LA a lot is the super charging time I know you mentioned you guys are working on supercharger 3 I assume it's going to be a bit faster just curious do you guys see room for kind of orders of magnitude improvement in charging time or are we kind of reaching a plateau with current battery chemistry and technology I wouldn't say that business order of magnitude improvement but I think a factor of 3 or
27:144 is is possible now the it won't be applicable to all battery chemistry's so 2012 chemistry can't take the charge rate of current chemistry but and worship could be we should good but but we just had to make a bunch of improvements to increase the charge rates like the key I think figure of Merit is that the ratio of drive time to charge time should be at least on the order of like six to one if not 10 eight to one or ten to one at the point of which you're driving say 10 times as much as you're charging then the natural sort of human need to take a break unless you have an enormous bladder is
28:02it's us to suspect our paramount if you start a road trip it's a 9 a.m. typically by around noon you want to stop you know it hit the restroom grab a bite to eat grab a coffee and and and that's you want to get back to your car and have it be ready to go so but that's where the it's like if you say that that's like maybe half an hour that's kind of like the minimum threshold for the car to be ready to go when you when you come back from a break and then what if you get to the point where it's they say ten to one where maybe it's only 15 20 minutes yeah something over 15 minutes then the car is ready to go away before you're ready to go and for some of the long-distance
28:44like if you get say a Model S 100 E and it could you can drive non-stop from LA to San Francisco if you drive carefully that's that's a long drive so and we think this your potential there certainly option a full range improvements down the road where will we hope to offer cars with ranges in excess of 400 miles hi my name is dr. Catherine
29:13vannacutt and I'm with People for the Ethical Treatment of Animals I have a question regarding your use of leather in your gear shifter and steering wheel gear shifter I guess if that dumping has any thing staring well maybe this or okay steering wheel so this but are you still using
29:40leather in some of your components so we do in Model S X and three currently we have the only leathers on the steering wheel rim and people have asked and kind of off menu we do replace that for people that are need a fully vegan car so our concern is that you know we're obviously facing an environmental crisis and animal agriculture as we all know is one of the the main contributors particularly leather production we're really pleased to see that you have introduced non leather seating options and that's a really fantastic step towards your goal of sustainability we
30:25would really like you to take the next step and eliminate all the other components yeah you know I think as X and 3 there we may be the first vehicles in production to kind of go non leather and at least in all of our seating in our trim and we're actively working on replacing the steering wheel as well we just want to make sure that the experience is as good if not better yeah just to add to that so there are some existing premium vegan leather suppliers ultra suede and Alcantara Ziff I'm pronouncing it correctly they used by other luxury cars like Ferrari so yeah we would really like to see in Tesla step up as well yeah well they're from offering so as process technically we
31:13don't say it on the website but you can actually have a Tesla that has zero leather whatsoever including on the steering wheel it's a little difficult because we do it in small quantities at the design studio so it's challenging to do it do it at scale but what model why for example will what would will not have any leather in it at including in the steering wheel even if it does
31:47thank you when do you expect significant battery advances to allow Tesla to pack twice as much energy into each of the batteries without increasing the size or weight twice as much as tricky but we can suddenly see a path to about a to about a 30 30 percent improvement maybe a 40 percent improvement in energy in the same size battery bank but like look the best technology we are confident does work and it just it needs to be scaled up and made very reliable but the 30 to
32:2940 percents definitely definitely doable long-term probably double long-term by other people's standards for Tesla standpoint reading probably 2 to 3 years to get to about a 30 percent improvement in volumetric energy density and yeah maybe six years or something six to eight years to get to a doubling this is highly dependent on making lithium for
33:00that really big jump a lithium anode is the key I just play it plating out pure lithium on the anode I started a company called Tesla attractions Tesla attractions is basically a gamification version of visiting superchargers I don't know that you tweeted about this a while ago saying that it was a good idea and if you ever decide to go through with it obviously I do not want to be in a position of competition with you so how could someone like me help the mission maybe we should interview Tesla
33:35you know a set focus right now is getting through the critical stuff like especially the model 3 production ramp and then smoothing out the production process to make the the car affordable getting the lower cost battery pack or smaller battery pack into production so those are kind of critical of fundamental things to really the survival of a company and then we do kind of like the fun fund provillus things like but there with a droid so the experience maybe later this year I think always Lyc doing sort of fun study things in the car like you know the Easter eggs that you know there's like a lot of Easter eggs in the Model S necks and and three that are quite fun if in
34:19fact once they discovered they're put in the Easter Egg box which you just tap the Tesla logo on the screen and and then wait for about 30 seconds and it opens the Easter Egg box yeah all right my name is pence Gerber I'm hooked into Tesla in multiple ways obviously I'm a stockholder I have a solo SolarCity panel on my roof and also have a model three reservation from March 31st of 2016 ready for configuring but I would really
34:53really really like to have an all-wheel drive if oh yeah can you give an order of magnitude guess as to how long I'd have to wait for that you should receive a configuration email within the next week or so yeah and we start we start production or wheel drive who actually take we were starting on production on overdrive this month and we expect to
35:20scale that up in that in July and August and and be in high-volume production of all-wheel drive by September for sure I have a tour planned for September one month two or all right well make sure you get any chance I can make it yes we'll make sure you get your car thank you hi I got my first model lesson 2013 and
35:48have an X and a three now I love the product I feel like I'll always be one of the first to try to try one of your new products and I was just wondering is there any chance in the future that we'll be able to text commands to our car like Pete my cart or this temperature come pick me up you know that type of stuff we're gonna keep enhancing the Tesla app on the phone and to be able to you know long time just tap the summon button and your car will come find you wherever you are and do it really wanted to learn what you see wheel you can change the temperature right now from the app but we want the car to learn what you most likely would do like it basically if
36:33there was a fit if there was a great show for in charge of the car what would that person do anticipating your needs and knowing what what you want so it's sort of like in an intuitive trusty steed that just always notice why you what you want ideally and then you can easily adjust that as needed so you won't just be as
37:01close like a mind-meld with the car as possible oh hi my name is Kevin I have the remote OS since 13 and my question is about auto pilot and the use of lidar is we all know that Tesla probably is the the company along they're not using lighter we're all other vendors using it so as autonomous driving is close to reality I think inevitably there will be a showdown on which approach will be most superior so what do you think when that time will be well
37:39I think I think lateral will be seen as the what light out tends to drive companies to do is to go to a local maximum in terms of autopilot capability or autonomous driving capability and light all ends up being like somewhat of a crutch it's it's so it's helpful to get almost there but if you if you rely on it you will never get actually get there is my opinion so you have to make vision work extremely well in order to
38:12achieve true self driving once you've made vision work extremely well lidar is really necessary it's not really adding anything we do have sophisticated sonar like ultrasonic sensors around the vehicle for near-field and we do have a forward radar system which is useful for detecting objects even in fog it's sort
38:40of snow/rain like low visibility conditions where you can't see what's going on and that's also a case where lidar is ineffective because ladder is an active active photon generator in the visible spectrum this doesn't make sense to me because you have a massive amount of incoming photons in the visible spectrum normally so if you're gonna do active photon generation the 400 to 700 nanometers is the wrong wavelength or on that orders wrong wrong wavelength you really want to be aiming for something that's around a four millimeter wavelength because that is a occlusion penetrating
39:25[Applause] we're two-time Tesla Model S owners and also stockholders question about JB every SEC filing as always included JP is a key man and he was suddenly dropped off in the q1 filing so a little concerned about JB oh really I know an accidental sorry I don't even know that was the case and is there a pallet away
40:00yeah did the intent was to make it more generic the risk factor there was no intended or unintended implication behind Jamie we love you watching sorry and one last one which is the Tesla's energy storage business is still running double-digit gross margins negative it looks like even though you've managed to deliver a lot more megawatt hours can you talk about the pathway to getting a profit out of that business please yeah
40:35sure yeah definitely we expect our storage business to turn significantly positive and our goal is to have the same level of gross margin as the automotive business and as our volume scale and as we have more power walls out there and our manufacturing efficiencies come in we definitely expect that you'll see a big change in a positive trend in that every quarter into 2019 they were raving forces
41:04essentially about the same gross margin level as in the cars which is sort of twenty to thirty percent gross margin for all of the energy products it's it's necessarily we're doing that as you ramp up production it ends up being it's negative just as was form all three and model snx but I probably later this year
41:25or certainly early next we should be in the 20 to 30% cost margin level for energy storage hi thanks for a shout on the conference call by the way and my question was just surrounding as a long-term investor I hate to say this but I feel like my trust in Tesla's timelines could sort of eroded a little bit with the model 3 ramp so should I keep discounting things on Elon time or you guys have you learned anything about this I think I do have like an issue with time it's like
41:59yeah it's been true since so my brother's here I was like I have a condition my brother used to like when we were catching the bus to school he would lie to me about the time and he always always says like or sometimes like earlier than it actually was and then I get there slightly after that and then we'd actually be able to catch the bus so I and you know this is something I'm trying to get better at a sort of
42:34naturally optimistic person which other which would not have probably done cars or rockets if I was not but so I'm trying to recalibrate these these estimates yeah try recalibrate much as possible yeah I I mean I've probably put some sandbag on future dates that's probably wise but I
43:08like I kind of today would say when I think it can occur but then I'm typically optimistic about these things but maybe less up hopefully less optimistic over time so yeah like it pretty much always happens but not exactly on the timeframe hey Alan we have we're about 15 minutes over one hour do you want to take maybe a couple more yeah maybe a couple more questions from each night okay hello mr. musk my
43:41name is Alex Perez I really like your jacket today and so it kind of got me thinking has Tesla ever thought about going into the motorcycle business yeah I actually used to ride motorcycles when I was a kid and like dirt biking for like eight years or something and then I had a road bike until I was 17
44:07and was almost killed by a truck so we're not gonna be motorcycles hi mr. musk my name is Sonia and my mom has a model three and my dad has a model s and I have one chair so I've noticed that whenever my mom is like kind of going a little bit too fast to like like approaching a car to stop sign the Tesla beeps pretty loudly to kind of warn like hey you're gonna hit that I was wondering like are you guys thinking about developing mode where you like drivers could choose to turn it on and then if a Tesla detected that it was approaching something quickly or was going to crash it would like gradually slow down a little bit so best-case scenario if the triberg takes control
44:51doesn't really matter worst-case scenario and they crash it decelerates so it's not as bad it does actually have that there's a automatic emergency braking and I think we're just things like instead of like last-minute kind of dramatic slowdown maybe slow down sooner but less dramatically and that that is that is something that will
45:22occur with the latest versions of autopilot so will if it will decrease speed proportionate to the the confidence level you would what we want to do that in a way that's not annoying to people like the car isn't slowing down a lot but it's really delicate balance between not annoying people so if they want to turn it off but also being safe for autopilot I
45:56think it's the improvements gonna be really quite quite dramatic in the over the next several months yeah that the system is is intended to change speed proportionate to its confidence in going forward but in order to do so we had to improve the sophistication of the order autopilot neural net and the heuristics that go with it so it just didn't annoy the hell out of people because there many times where the car thinks it should slow down but actually not really and and I would just drive people crazy so I do you think what you're getting at is something that you will see play out with the versions of autopilot that are deploying later this year and including
46:45the one that's coming out this week hi guys I just wanted to personally thank Neil on JB Franz Deepak could have you back Tim Emily Vinny Andrew the people at Tesla who I don't know by name the proud owner of an early 2012 Model S very different than model 3 that my dad recently got his first new car he's pretty perplexed and amazed by it I actually drove that from New Jersey to Richmond probably could have skipped supercharging entirely charged the car for I think it was maybe about 20 minutes and continued the trip and the car was actually telling me that I was ready to go before I was done eating the sandwich I picked up oh I just wanted to ask a quick question
47:34it's our possibility get a business card for the guiity or left because I wanted to tell him about a company this man's been working on disruptive technology that you're already indirectly using and wanted an opportunity to speak with you about it further Thanks all right thanks yeah we'll certainly take the business card and always looking to investigate interesting opportunities to improve the car if yeah okay you know one thing I think we haven't been good at educating people on is that the Model S and X but especially the Model S has improved very dramatically from 2012 so we're arguably on version three or four of the Model S right now so it's really a gigantic
48:20improvement in Model S today versus Model S of 2012 when your first auto production I really encourage anyone who's gone to early Model S to test drive the current version I think you'll be blown away by how how much it's improved alright I think that might be the last question right I will soon be driving around in Model 3 curious - you mentioned earlier that during production it's people first in the automation later slowly what kind of considerations prompted to
48:55do otherwise when you started the model 3 production can you talk about that I think we're just overconfident about the degree of automation that was possible and we we did rely quite a bit on tier 1 or automated manufacturing automation integrators and a couple of those things really didn't work out at all and now we are we're really going to internalize all Tier one manufacturing systems at
49:28Tesla so we'll have a lot of suppliers but they'll be at the tier 2 and tier 3 level alright thanks very much for coming [Music]
资源与出处
- 视频在 YouTube 收看
- 档案silicon-valley.tech 档案页本页视频与文稿数据来自 Elon Musk Archive
- 精读本篇精读 Markdown 版