BV1CWXvBME1K · 浑水摸鱼清源1 00:00:00,140 --> 00:00:05,059 我是塔莎·金尼，我是ARC和研究部门的投资分析主管 2 00:00:05,059 --> 00:00:08,269 负责自动驾驶技术与机器人团队 3 00:00:08,269 --> 00:00:11,808 今天我将为大家介绍自动驾驶汽车 4 00:00:11,808 --> 00:00:15,929 如今无人驾驶出租车已在部分地区向公众开放 5 00:00:15,929 --> 00:00:16,908 在选定城市 6 00:00:16,908 --> 00:00:17,449 在美国 7 00:00:17,449 --> 00:00:20,730 在中国和中东地区 8 00:00:20,730 --> 00:00:21,910 嗯 9 00:00:21,910 --> 00:00:23,850 你们知道我们过去常被问到 10 00:00:23,850 --> 00:00:25,870 这项技术现在是否可行 11 00:00:25,870 --> 00:00:26,109 现在 12 00:00:26,109 --> 00:00:27,129 这已经很清楚了 13 00:00:27,129 --> 00:00:29,410 有一辆车正在四处行驶 14 00:00:29,410 --> 00:00:34,518 无需前排或乘客座位有人，能够完全自主载客 15 00:00:34,518 --> 00:00:37,118 有趣且在过去一年发展的是 16 00:00:37,118 --> 00:00:40,490 在运营区域内 17 00:00:40,490 --> 00:00:43,850 Waymo实际上正在与Uber和Lyft竞争 18 00:00:43,850 --> 00:00:45,270 尤其是这里的Lyft 19 00:00:45,270 --> 00:00:48,018 从市场份额来看 20 00:00:48,018 --> 00:00:51,659 尽管车队规模相对较小 21 00:00:51,659 --> 00:00:54,420 但已提供具有竞争力的服务 22 00:00:54,420 --> 00:00:58,579 与当前的人类驾驶乘车选项相比 23 00:00:58,679 --> 00:00:59,198 嗯 24 00:00:59,198 --> 00:01:02,179 但正如我提到的车队规模需扩大 25 00:01:02,179 --> 00:01:05,189 我们认为数据优势仍是关键 26 00:01:05,189 --> 00:01:07,049 这里有两张图表 27 00:01:07,049 --> 00:01:10,870 由我的同事丹尼尔·麦奎尔制作 28 00:01:10,870 --> 00:01:12,049 无人驾驶出租车里程 29 00:01:12,049 --> 00:01:17,030 这是实际商业运营的无人驾驶出租车服务数据 30 00:01:17,030 --> 00:01:18,299 可以看到这里 31 00:01:18,299 --> 00:01:22,239 Waymo在美国和百度的Apollo项目在中国 32 00:01:22,239 --> 00:01:26,260 在商业无人驾驶里程方面领先 33 00:01:26,260 --> 00:01:30,439 但当我们查看总数据采集量时 34 00:01:30,439 --> 00:01:31,500 即这张图表 35 00:01:31,500 --> 00:01:33,629 最右侧的图表 36 00:01:33,629 --> 00:01:34,269 嗯 37 00:01:34,269 --> 00:01:37,079 这是累计自动驾驶里程 38 00:01:37,079 --> 00:01:37,439 嗯 39 00:01:37,439 --> 00:01:43,920 可以看到特斯拉在全球遥遥领先其他竞争对手 40 00:01:43,920 --> 00:01:49,540 这是因为他们收集了客户车辆的视频数据 41 00:01:49,540 --> 00:01:52,579 所有车辆都配备了传感器组 42 00:01:52,579 --> 00:01:56,299 具备传感器组 43 00:01:56,299 --> 00:01:56,560 并且 44 00:01:56,560 --> 00:01:59,049 并搭载了车载计算系统 45 00:01:59,049 --> 00:02:03,349 能够通过软件升级将车辆变为完全自动驾驶 46 00:02:03,379 --> 00:02:07,609 因此特斯拉拥有无可比拟的数据优势 47 00:02:07,609 --> 00:02:08,639 无人能及 48 00:02:08,639 --> 00:02:09,598 这得益于 49 00:02:09,598 --> 00:02:13,318 他们长期拥有具备该功能的车辆 50 00:02:13,318 --> 00:02:15,620 下线即具备此能力 51 00:02:15,620 --> 00:02:16,039 嗯 52 00:02:16,039 --> 00:02:18,520 再加上他们拥有庞大的 53 00:02:18,520 --> 00:02:19,039 呃 54 00:02:19,039 --> 00:02:20,780 消费者车队正在推动这一 55 00:02:20,780 --> 00:02:21,400 因此他们 56 00:02:21,400 --> 00:02:21,759 你知道的 57 00:02:21,759 --> 00:02:23,159 数百万辆车与 58 00:02:23,159 --> 00:02:23,639 呃 59 00:02:23,639 --> 00:02:25,629 竞争对手的数千辆 60 00:02:25,629 --> 00:02:31,270 所以我们认为数据优势是规模化这项服务并推动技术发展的关键 61 00:02:31,270 --> 00:02:31,770 呃 62 00:02:31,770 --> 00:02:34,189 未来的技术发展 63 00:02:34,680 --> 00:02:38,699 那么我们为什么在讨论自动驾驶出租车呢 嗯 64 00:02:38,699 --> 00:02:43,379 这里最关键的要点是 65 00:02:43,379 --> 00:02:45,340 我们多年来一直在强调 66 00:02:45,340 --> 00:02:50,139 自动驾驶汽车将比人类驾驶更便宜 67 00:02:50,139 --> 00:02:52,210 我的意思是 68 00:02:52,210 --> 00:02:52,650 嗯 69 00:02:52,650 --> 00:02:54,330 如果你看右边的图表 70 00:02:54,330 --> 00:02:54,810 可以看到 71 00:02:54,810 --> 00:03:00,569 西方市场的人类驾驶乘车费用平均略低于三美元 72 00:03:00,569 --> 00:03:01,680 考虑通胀因素 73 00:03:01,680 --> 00:03:02,280 嗯 74 00:03:02,280 --> 00:03:06,569 而个人驾车成本略低于一美元 75 00:03:06,569 --> 00:03:13,849 自动驾驶出租车规模化后可定价为消费者25美分 76 00:03:13,849 --> 00:03:14,550 嗯 77 00:03:14,550 --> 00:03:18,889 当然中间有很多价格区间在2美元到80美分之间 78 00:03:18,889 --> 00:03:22,689 甚至可能高于2.8美元的高端市场 79 00:03:22,689 --> 00:03:23,159 比如 80 00:03:23,159 --> 00:03:23,840 嗯 81 00:03:23,840 --> 00:03:24,879 但 82 00:03:24,879 --> 00:03:25,319 你知道的 83 00:03:25,319 --> 00:03:27,699 在2.8美元和25美分之间 84 00:03:27,699 --> 00:03:30,960 其实有很多支持更高价位的需求 85 00:03:30,960 --> 00:03:33,759 我们认为自动驾驶极具吸引力 86 00:03:33,759 --> 00:03:36,259 因为它将降低交通成本 87 00:03:36,259 --> 00:03:38,659 比你开私家车更便宜 88 00:03:38,659 --> 00:03:40,889 比打车更便宜 89 00:03:40,889 --> 00:03:45,169 但规模化后可能低至25美分 90 00:03:45,169 --> 00:03:48,930 而在早期阶段最关键的是 91 00:03:48,930 --> 00:03:49,629 呃 92 00:03:49,629 --> 00:03:52,650 正如我之前提到的规模化 93 00:03:52,650 --> 00:03:53,250 嗯 94 00:03:53,250 --> 00:03:54,689 特斯拉的情况 95 00:03:54,689 --> 00:03:58,599 在制造规模化方面同样至关重要 96 00:03:58,599 --> 00:03:58,960 嗯 97 00:03:58,960 --> 00:04:00,020 你可以在这里看到 98 00:04:00,020 --> 00:04:00,539 呃 99 00:04:00,539 --> 00:04:03,050 左边的图表显示 100 00:04:03,050 --> 00:04:05,490 所以比较Waymo和特斯拉 101 00:04:05,490 --> 00:04:05,969 你知道吗 102 00:04:05,969 --> 00:04:08,669 Waymo从汽车制造商采购车辆 103 00:04:08,669 --> 00:04:10,449 然后进行改装 104 00:04:10,449 --> 00:04:12,650 他们与合作伙伴进行整合 105 00:04:12,650 --> 00:04:14,759 在车辆上安装传感器组 106 00:04:14,759 --> 00:04:15,259 嗯 107 00:04:15,259 --> 00:04:15,960 所以如果你 108 00:04:15,960 --> 00:04:16,879 如果你看看这个 109 00:04:16,879 --> 00:04:18,100 你知道这辆车的价格 110 00:04:18,100 --> 00:04:19,220 加上传感器组 111 00:04:19,220 --> 00:04:19,740 而 112 00:04:19,740 --> 00:04:23,699 当然会增加车辆的折旧成本 113 00:04:23,699 --> 00:04:27,569 因为它比特斯拉生产线上的普通车型更昂贵 114 00:04:27,569 --> 00:04:28,230 嗯 115 00:04:28,230 --> 00:04:32,500 我们认为今天道路上行驶的车辆 116 00:04:32,500 --> 00:04:35,199 Waymo的上一代车型和Model系列 117 00:04:35,199 --> 00:04:38,279 特斯拉已在其自动驾驶车队中使用 118 00:04:38,279 --> 00:04:41,600 特斯拉已经具备35%的成本优势 119 00:04:41,600 --> 00:04:47,779 但规模化后每英里成本优势可能达50% 120 00:04:47,779 --> 00:04:50,959 这对早期商业化至关重要 121 00:04:50,959 --> 00:04:53,540 当车队利用率较低时 122 00:04:53,540 --> 00:04:57,019 车辆的基础成本影响显著 123 00:04:57,019 --> 00:04:58,110 然后你知道 124 00:04:58,110 --> 00:05:02,699 规模化后预计车辆利用率 125 00:05:02,699 --> 00:05:05,560 即车辆实际行驶频率 126 00:05:05,560 --> 00:05:09,100 通过产生收入来降低每英里成本 127 00:05:10,100 --> 00:05:10,839 但 128 00:05:10,839 --> 00:05:11,300 你知道 129 00:05:11,300 --> 00:05:12,819 这说明 130 00:05:12,819 --> 00:05:15,899 垂直整合的汽车制造商更具优势 131 00:05:17,389 --> 00:05:17,910 你知道 132 00:05:17,910 --> 00:05:20,790 自动驾驶出行或主导城市交通 133 00:05:20,790 --> 00:05:25,470 我的同事丹尼尔今年早些时候发布了一篇精彩博客 134 00:05:25,470 --> 00:05:29,269 分析了这些主要城市高峰时段 135 00:05:29,269 --> 00:05:35,689 可以看到人们日常出行的高峰与低谷 136 00:05:35,689 --> 00:05:36,350 嗯 137 00:05:36,350 --> 00:05:38,949 我们想强调的核心观点是 138 00:05:38,949 --> 00:05:41,600 如果观察当前城市里程 139 00:05:41,600 --> 00:05:45,319 网约车仅占其中很小比例 140 00:05:45,319 --> 00:05:47,658 约1% 141 00:05:48,059 --> 00:05:51,158 尽管某些城市可能高达10% 142 00:05:51,158 --> 00:05:53,019 在美国约为1% 143 00:05:53,019 --> 00:05:53,779 嗯 144 00:05:53,779 --> 00:05:55,600 自动驾驶出租车 145 00:05:55,600 --> 00:05:57,160 需要更少 146 00:05:57,160 --> 00:05:57,680 你知道 147 00:05:57,680 --> 00:06:01,500 更少的自动驾驶出租车可替代现有车辆 148 00:06:01,500 --> 00:06:07,480 只需替代当前的人类驾驶网约车 149 00:06:07,480 --> 00:06:10,379 仅需14万辆自动驾驶出租车 150 00:06:10,379 --> 00:06:16,250 因为它们的利用率将高于现有网约车 151 00:06:16,250 --> 00:06:16,670 嗯 152 00:06:16,670 --> 00:06:18,189 你需要从这个角度理解 153 00:06:18,189 --> 00:06:18,850 这就像 154 00:06:18,850 --> 00:06:22,269 可能相当于特斯拉一个月的产量 155 00:06:22,269 --> 00:06:24,110 你可以从图表中看到 156 00:06:24,110 --> 00:06:24,529 呃 157 00:06:24,529 --> 00:06:25,649 特斯拉实际上通过 158 00:06:25,649 --> 00:06:26,089 呃 159 00:06:26,089 --> 00:06:28,129 你知道他们现有的车型阵容 160 00:06:28,129 --> 00:06:29,769 当然，如果你看整体数据 161 00:06:29,769 --> 00:06:32,360 Cyber Cab预期产能 162 00:06:32,360 --> 00:06:34,470 他们已经能够满足 163 00:06:34,470 --> 00:06:35,769 嗯 嗯 164 00:06:35,769 --> 00:06:37,970 在许多主要城市的出行需求 165 00:06:37,970 --> 00:06:38,490 就在这里 166 00:06:38,490 --> 00:06:40,209 美国主要山城地区 167 00:06:40,209 --> 00:06:42,069 就美国整体而言 168 00:06:42,069 --> 00:06:43,610 两到四百万辆自动驾驶出租车 169 00:06:43,610 --> 00:06:47,990 实际上仅占美国注册车辆的不到10% 170 00:06:47,990 --> 00:06:48,410 嗯 171 00:06:48,410 --> 00:06:51,189 凭借更高的使用率 172 00:06:51,189 --> 00:06:53,709 实际上可以覆盖所有城市里程 173 00:06:54,370 --> 00:06:57,069 自动驾驶出租车可能产生 174 00:06:57,069 --> 00:07:01,918 到2030年约34万亿美元的企业价值 175 00:07:01,918 --> 00:07:03,439 我们之前曾提到过 176 00:07:03,439 --> 00:07:04,619 我再重复一遍 177 00:07:04,619 --> 00:07:06,839 这里我们认为 178 00:07:06,839 --> 00:07:10,860 大部分经济收益将流向自动驾驶技术提供商 179 00:07:10,860 --> 00:07:12,000 平台提供商 180 00:07:12,000 --> 00:07:14,809 这些公司构建了相关技术 181 00:07:14,809 --> 00:07:17,689 使车辆能够自动驾驶的技术栈 182 00:07:17,689 --> 00:07:20,939 因为这才是降低每英里成本的关键 183 00:07:20,939 --> 00:07:22,060 而最终驱动需求 184 00:07:22,060 --> 00:07:24,379 我们将看到这一点 185 00:07:24,379 --> 00:07:27,160 这个预测的最大风险 186 00:07:27,160 --> 00:07:30,538 我要指出的是除了特斯拉 187 00:07:30,538 --> 00:07:35,670 许多传统车企及其技术合作伙伴如Waymo 188 00:07:35,670 --> 00:07:36,230 嗯 189 00:07:36,230 --> 00:07:41,589 尚未做出大规模的自动驾驶车队承诺 190 00:07:44,610 --> 00:07:46,910 所以我们在此做了一定假设 191 00:07:46,910 --> 00:07:47,310 你知道的 192 00:07:47,310 --> 00:07:47,810 我们确实 193 00:07:47,810 --> 00:07:48,209 呃 194 00:07:48,209 --> 00:07:49,170 中国有许多参与者 195 00:07:49,170 --> 00:07:49,790 例如 196 00:07:49,790 --> 00:07:51,709 事情可以迅速推进 197 00:07:51,709 --> 00:07:52,629 几乎一夜之间 198 00:07:52,629 --> 00:07:54,339 在制造规模方面 199 00:07:54,339 --> 00:07:57,279 正在着手构建自动驾驶出租车平台 200 00:07:57,279 --> 00:07:59,779 因此我们预计会有更多竞争对手涌现 201 00:07:59,779 --> 00:08:01,040 嗯，但我要指出的是 202 00:08:01,040 --> 00:08:03,889 这实际上是对今天预测的最大风险 203 00:08:04,250 --> 00:08:05,269 最后一点是 204 00:08:05,269 --> 00:08:05,569 嗯 205 00:08:05,569 --> 00:08:05,769 你知道的 206 00:08:05,769 --> 00:08:06,689 我们在这里阐述 207 00:08:06,689 --> 00:08:07,798 那个呃 208 00:08:07,798 --> 00:08:10,759 生态系统中的不同组成部分 209 00:08:10,759 --> 00:08:11,218 嗯 210 00:08:11,218 --> 00:08:12,038 在之前的幻灯片中 211 00:08:12,038 --> 00:08:14,928 我们概述了一些 212 00:08:14,928 --> 00:08:18,108 它将为每个部分带来的价值再次 213 00:08:18,108 --> 00:08:18,728 那个呃 214 00:08:18,728 --> 00:08:21,930 拼图中需要占据的关键部分以获取 215 00:08:21,930 --> 00:08:26,250 大部分经济收益来自自动驾驶技术平台部分 216 00:08:26,889 --> 00:08:29,550 我们认为完全电动化 217 00:08:29,550 --> 00:08:30,089 嗯 218 00:08:30,089 --> 00:08:31,129 汽车制造商 219 00:08:31,129 --> 00:08:31,990 嗯 220 00:08:31,990 --> 00:08:33,409 在这里拥有发展空间 221 00:08:33,409 --> 00:08:34,389 嗯我们 222 00:08:34,389 --> 00:08:34,730 你知道的 223 00:08:34,730 --> 00:08:37,068 大约不到10%的收入 224 00:08:37,068 --> 00:08:39,208 约占企业价值的1% 225 00:08:39,208 --> 00:08:40,509 这实际上是因为再次 226 00:08:40,509 --> 00:08:43,509 我们看到的平台侧企业价值规模更大 227 00:08:43,509 --> 00:08:46,509 然后是车队运营商或领航者 228 00:08:46,509 --> 00:08:51,809 这些公司如RideHill Partners等科技供应商选择的 229 00:08:51,809 --> 00:08:55,149 以及负责维护车辆的公司 230 00:08:55,149 --> 00:08:55,629 嗯 231 00:08:55,629 --> 00:08:57,870 这些可能不是我们今天熟知的公司 232 00:08:57,870 --> 00:08:58,269 顺便说一句 233 00:08:58,269 --> 00:09:00,529 可能是该领域新兴的初创企业 234 00:09:00,529 --> 00:09:04,870 这也是启动自动驾驶出租车服务所需的重要拼图部分 235 00:09:04,870 --> 00:09:08,340 当然有些参与者可能同时属于三大类别 236 00:09:08,340 --> 00:09:10,019 你可以在这里看到我们划分的 237 00:09:10,019 --> 00:09:11,210 这部分内容在幻灯片中 238 00:09:11,210 --> 00:09:11,789 嗯 239 00:09:11,789 --> 00:09:13,330 但再次强调 240 00:09:13,330 --> 00:09:16,970 创造最大价值的领域是技术平台
评论采集失败
采集完成，该视频无弹幕