BV1XEML69E4i · 浑水摸鱼清源1 00:00:00,280 --> 00:00:02,160 大家好，欢迎回到今天的视频 2 00:00:02,160 --> 00:00:02,759 在今天的视频中 3 00:00:02,759 --> 00:00:03,879 我们来谈谈元宇宙 4 00:00:03,879 --> 00:00:06,839 同时也要关注更广泛的AI领域 5 00:00:06,839 --> 00:00:11,849 因为目前市场正经历非常有趣的波动 6 00:00:11,849 --> 00:00:12,968 正如你所看到的 7 00:00:12,968 --> 00:00:14,820 盘面出现了大量下跌 8 00:00:14,820 --> 00:00:16,679 但在这波红海行情中 9 00:00:16,679 --> 00:00:19,920 尤其是像谷歌或微软这样的科技巨头 10 00:00:19,920 --> 00:00:21,980 Meta和Palantir股价小幅下跌 11 00:00:21,980 --> 00:00:25,710 今日下跌3.8%，但我们仍有一些上涨个股 12 00:00:25,710 --> 00:00:29,969 特别是在半导体和新云领域 13 00:00:29,969 --> 00:00:32,399 观察新云领域相关公司 14 00:00:32,399 --> 00:00:33,380 Cory股价上涨 15 00:00:33,380 --> 00:00:35,299 Nebues和Core Scientific上涨 16 00:00:35,299 --> 00:00:36,159 I Iran 17 00:00:36,159 --> 00:00:37,320 今日全部上涨 18 00:00:37,320 --> 00:00:40,429 当然昨日它们大幅下跌 19 00:00:40,429 --> 00:00:45,009 但半导体板块今日也上涨 20 00:00:45,090 --> 00:00:48,009 所以为何出现这种分化 21 00:00:48,009 --> 00:00:48,250 没错 22 00:00:48,250 --> 00:00:49,689 因为科技巨头普遍下跌 23 00:00:49,689 --> 00:00:52,920 这又是一个故事 24 00:00:52,920 --> 00:00:56,479 预计AI资本支出将大幅增加 25 00:00:56,479 --> 00:00:59,320 因此科技巨头股价承压 26 00:00:59,320 --> 00:01:03,509 这也是半导体和新云板块回升的原因 27 00:01:03,509 --> 00:01:04,750 我同意这个观点 28 00:01:04,750 --> 00:01:07,009 这也是原因之一 29 00:01:07,009 --> 00:01:12,879 但今日头条无疑聚焦半导体分析 30 00:01:12,879 --> 00:01:15,859 当然，每当半导体分析发布看空报告 31 00:01:15,859 --> 00:01:17,219 人们并不信任他们 32 00:01:17,219 --> 00:01:18,939 但每次他们发布看多观点 33 00:01:18,939 --> 00:01:21,618 我们自然会跟风炒作 34 00:01:21,618 --> 00:01:24,078 毕竟我们只想听利好消息 35 00:01:24,078 --> 00:01:25,689 这确实合理 36 00:01:25,689 --> 00:01:27,290 不不 37 00:01:27,290 --> 00:01:29,269 先不继续这个话题 38 00:01:29,269 --> 00:01:31,129 当然还要关注金融科技板块 39 00:01:31,129 --> 00:01:34,290 今日金融科技板块全线下跌 40 00:01:34,290 --> 00:01:40,120 尽管这些个股近期有所反弹 41 00:01:40,120 --> 00:01:43,290 现在可能进入短暂回调 42 00:01:43,290 --> 00:01:44,969 毕竟美国输给比利时 43 00:01:44,969 --> 00:01:48,750 因此他们决定回归中东事务 44 00:01:48,750 --> 00:01:51,840 或许不想再次被嘲笑 45 00:01:51,840 --> 00:01:53,019 再次来看 46 00:01:53,019 --> 00:01:54,739 可能油价上涨 47 00:01:54,739 --> 00:01:56,760 通胀持续高位 48 00:01:56,760 --> 00:01:57,859 无需降息 49 00:01:57,859 --> 00:02:02,090 或加息对这些板块造成负面影响 50 00:02:02,090 --> 00:02:02,670 如你所知 51 00:02:02,670 --> 00:02:04,129 我喜欢谈论元宇宙 52 00:02:04,129 --> 00:02:05,870 我喜欢谈论元宇宙 53 00:02:05,870 --> 00:02:06,469 非常多 54 00:02:06,469 --> 00:02:09,449 所以我无法对元宇宙股票保持沉默 55 00:02:09,449 --> 00:02:13,370 就像一年前我无法对谷歌保持沉默一样 56 00:02:13,370 --> 00:02:14,409 关于元宇宙 57 00:02:14,409 --> 00:02:15,370 当然你知道的 58 00:02:15,370 --> 00:02:18,139 我们稍后会深入讨论所有新闻和公告 59 00:02:18,139 --> 00:02:21,259 我们知道股价大幅下跌的原因 60 00:02:21,259 --> 00:02:24,060 但我们不清楚为何需要这么便宜 61 00:02:24,060 --> 00:02:25,860 但公司仍然非常盈利 62 00:02:25,860 --> 00:02:27,120 这是一家增长迅速的好公司 63 00:02:27,120 --> 00:02:27,979 等等等等 64 00:02:27,979 --> 00:02:32,659 所以资本支出用于AI，这就是主要原因之一 65 00:02:32,659 --> 00:02:35,479 我们不清楚长期投资回报率会如何 66 00:02:35,479 --> 00:02:38,379 因此市场并不看好这一情况 67 00:02:38,379 --> 00:02:41,240 其次还有诉讼问题 68 00:02:41,240 --> 00:02:44,689 目前正在审理的青少年诉讼案 69 00:02:45,090 --> 00:02:49,469 新墨西哥州几个月前的处罚金额是多少 70 00:02:49,469 --> 00:02:54,909 据我所知他们需支付约两到三亿美元罚款 71 00:02:54,909 --> 00:02:56,629 据报道 72 00:02:56,629 --> 00:02:56,830 哦 73 00:02:56,830 --> 00:03:02,020 他们可能面临高达1400亿或1500亿美元的罚款 74 00:03:02,020 --> 00:03:04,840 他们的市值已达1500亿美元 75 00:03:04,840 --> 00:03:06,219 所以我不确定 76 00:03:06,219 --> 00:03:07,960 可能只是耸人听闻的标题 77 00:03:07,960 --> 00:03:11,229 但我不认为这会成为现实 78 00:03:11,229 --> 00:03:12,189 不过好吧 79 00:03:12,189 --> 00:03:15,389 这些都是影响股价的因素 80 00:03:15,389 --> 00:03:16,509 比如元宇宙 81 00:03:16,509 --> 00:03:19,550 但但首先是 82 00:03:19,550 --> 00:03:20,129 顺便说一句 83 00:03:20,129 --> 00:03:20,729 第二部分 84 00:03:20,729 --> 00:03:21,289 这些诉讼 85 00:03:21,289 --> 00:03:22,949 我认为诉讼将于八月开始 86 00:03:22,949 --> 00:03:23,969 或者类似时间 87 00:03:23,969 --> 00:03:28,800 所以我们未来一个月左右会获得更多信息 88 00:03:28,800 --> 00:03:30,560 在资本支出方面 89 00:03:30,560 --> 00:03:31,300 在AI领域 90 00:03:31,300 --> 00:03:32,159 这方面 91 00:03:32,159 --> 00:03:35,930 我们现在获得更多信息 92 00:03:35,930 --> 00:03:38,370 我们也会收到更多公告 93 00:03:38,370 --> 00:03:40,889 随着财报发布 94 00:03:40,889 --> 00:03:42,550 大约三周后 95 00:03:42,550 --> 00:03:46,020 公司将提供更多答案 96 00:03:46,020 --> 00:03:47,500 而昨天 97 00:03:47,500 --> 00:03:48,360 昨天 98 00:03:48,360 --> 00:03:49,479 当我录制这段时 99 00:03:49,479 --> 00:03:51,939 他们推出了Muse图像和Muse视频 100 00:03:51,939 --> 00:03:56,590 当然，超级智能团队首个大模型是Mu Spark 101 00:03:56,590 --> 00:03:58,870 那已经是两三个月前的事了 102 00:03:58,870 --> 00:04:01,490 我们很快应该会收到新更新 103 00:04:01,490 --> 00:04:03,959 但今年这是一个非常明智的举措 104 00:04:03,959 --> 00:04:05,498 使用图像使用视频 105 00:04:05,498 --> 00:04:07,938 为什么因为这很有道理 你有Instagram 106 00:04:07,938 --> 00:04:09,479 Instagram上有数十亿用户 107 00:04:09,479 --> 00:04:11,909 Facebook上有数十亿用户 108 00:04:11,909 --> 00:04:12,870 在Threads上 109 00:04:12,870 --> 00:04:14,789 你有数亿用户 110 00:04:14,789 --> 00:04:23,410 但图像视频对我来说始终是 他是否要变现你的AI方案 111 00:04:23,410 --> 00:04:26,529 尤其是现在有了非常好的模型 112 00:04:26,529 --> 00:04:28,560 如果你拥有非常好的模型 113 00:04:28,560 --> 00:04:32,459 并且拥有分发渠道和广告资源 114 00:04:32,459 --> 00:04:37,589 并且还在与中小企业等合作等等 115 00:04:37,589 --> 00:04:39,790 你可以开始升级这些服务 116 00:04:39,790 --> 00:04:42,529 你可以开始变现AI 117 00:04:42,529 --> 00:04:45,189 这是市场需要的东西 118 00:04:45,189 --> 00:04:48,029 我认为这就是我们将要看到的 119 00:04:48,029 --> 00:04:51,829 随着Muse Image和Muse Video的发布 120 00:04:51,829 --> 00:04:53,750 稍后我会详细解释 121 00:04:53,750 --> 00:04:56,129 接下来的重大宣布就是这个 122 00:04:56,129 --> 00:04:57,089 其实并不是正式宣布 123 00:04:57,089 --> 00:04:58,750 只是半分析报告 124 00:04:58,750 --> 00:05:04,350 《Tropic Q三2026年利润超十亿美元》 125 00:05:04,350 --> 00:05:07,069 《熵科技IPO财务前瞻》 126 00:05:07,069 --> 00:05:10,860 当然Tropic和OPI即将上市 127 00:05:10,860 --> 00:05:13,399 我们可以深入了解这些公司 128 00:05:13,399 --> 00:05:18,689 可以看到它们实际增长速度和盈利能力 129 00:05:18,689 --> 00:05:20,949 我迫不及待想看到它们上市 130 00:05:20,949 --> 00:05:25,810 因为那时我们终于能得到答案 减少疑问 131 00:05:25,810 --> 00:05:34,860 当然熵科技在2026年已有超十亿美元利润 132 00:05:35,238 --> 00:05:38,238 这非常看涨 133 00:05:38,238 --> 00:05:41,418 整个AI故事因为这证明了确实 134 00:05:41,418 --> 00:05:43,788 看来投入是合理的 135 00:05:43,788 --> 00:05:47,439 如果已能产生十亿美元利润 136 00:05:47,439 --> 00:05:50,660 接下来是估值问题 137 00:05:50,660 --> 00:05:51,399 毕竟 138 00:05:51,399 --> 00:05:53,939 十亿美元利润确实很可观 139 00:05:53,939 --> 00:05:56,389 增长速度极快 140 00:05:56,550 --> 00:05:59,779 但是否应值万亿市值 141 00:06:00,180 --> 00:06:03,129 我的回答大概是否定的 142 00:06:03,129 --> 00:06:05,910 但回到Meta时我又想 143 00:06:05,910 --> 00:06:06,350 好的 144 00:06:06,350 --> 00:06:09,290 熵科技已有十亿美元利润 145 00:06:09,290 --> 00:06:12,230 显然估值约万亿级别 146 00:06:12,230 --> 00:06:13,050 哦 147 00:06:13,050 --> 00:06:15,610 为何Meta仅值一点五万亿 148 00:06:15,610 --> 00:06:18,620 这完全不合逻辑 149 00:06:18,620 --> 00:06:22,389 有分析称被低估二十倍 150 00:06:22,389 --> 00:06:27,269 预计2027或2028年ARR达二十亿美元 151 00:06:27,269 --> 00:06:28,589 或者类似这样的数字 152 00:06:28,589 --> 00:06:31,579 他们的估值达到了六万亿美元 153 00:06:31,579 --> 00:06:34,420 这将使他们成为全球最有价值的公司 154 00:06:35,540 --> 00:06:37,060 这合理吗 155 00:06:37,339 --> 00:06:39,589 可能不太可能 156 00:06:39,589 --> 00:06:41,029 这就是现在的情况 157 00:06:41,029 --> 00:06:42,490 在整个AI故事中 158 00:06:42,490 --> 00:06:43,389 我感觉 159 00:06:43,389 --> 00:06:43,759 哦 160 00:06:43,759 --> 00:06:45,240 你是一家私营公司 161 00:06:45,240 --> 00:06:47,459 也许不仅限于AI，比如蓝色起源 162 00:06:47,459 --> 00:06:48,100 例如 163 00:06:48,100 --> 00:06:50,430 他们以130亿美元的估值融资 164 00:06:50,430 --> 00:06:53,470 我相信是130亿美元的融资前估值 165 00:06:53,470 --> 00:06:53,829 对的 166 00:06:53,829 --> 00:06:55,029 如果我没记错的话 167 00:06:55,029 --> 00:06:56,269 所以人们会自动说 168 00:06:56,269 --> 00:06:56,490 哦 169 00:06:56,490 --> 00:06:59,088 蓝色起源价值130亿美元 170 00:06:59,088 --> 00:07:00,668 火箭实验室价值50亿美元 171 00:07:00,668 --> 00:07:03,488 因此火箭实验室也应价值100亿美元 172 00:07:03,488 --> 00:07:06,060 没有人能被严重高估 173 00:07:06,060 --> 00:07:09,079 并不意味着其他公司也应被严重高估 174 00:07:09,079 --> 00:07:09,920 同样适用于OpenAI 175 00:07:09,920 --> 00:07:11,439 同样适用于Entropic 176 00:07:11,439 --> 00:07:13,680 也许数据会破裂或类似情况 177 00:07:13,908 --> 00:07:17,668 私营公司通常被严重高估 178 00:07:17,668 --> 00:07:19,649 尤其是即将上市时 179 00:07:19,649 --> 00:07:21,470 但这并不意味着 180 00:07:21,470 --> 00:07:21,709 哦 181 00:07:21,709 --> 00:07:26,050 所有公司都应该按50倍营收或30倍营收交易 182 00:07:26,050 --> 00:07:27,089 或者什么的 183 00:07:27,089 --> 00:07:31,189 可能意味着我们将看到这些股票大幅回调 184 00:07:31,189 --> 00:07:34,069 以及已上市公司 185 00:07:34,069 --> 00:07:35,129 无论是微软 186 00:07:35,129 --> 00:07:36,170 无论是Meta 187 00:07:36,170 --> 00:07:37,389 无论是亚马逊 188 00:07:37,389 --> 00:07:38,209 随便哪个 189 00:07:38,209 --> 00:07:42,399 在我看来这些都被严重低估 190 00:07:42,399 --> 00:07:43,139 所以是的 191 00:07:43,139 --> 00:07:44,079 在某个时候 192 00:07:44,079 --> 00:07:47,089 我认为这种脱节会破裂 193 00:07:47,089 --> 00:07:49,949 我认为如果这句话有道理的话 194 00:07:49,949 --> 00:07:51,470 我认为市场会说 195 00:07:51,470 --> 00:07:51,670 哦 196 00:07:51,670 --> 00:07:52,449 你知道吗 197 00:07:52,449 --> 00:07:53,329 这说不通 198 00:07:53,329 --> 00:07:57,000 Meta、微软、亚马逊、英伟达 199 00:07:57,000 --> 00:08:00,600 这类股票以极低市盈率交易 200 00:08:00,600 --> 00:08:05,329 而我们却给那些盈利不佳的公司估值 201 00:08:05,329 --> 00:08:07,769 对我来说保费很高 202 00:08:07,769 --> 00:08:09,529 现在这完全说不通 203 00:08:09,529 --> 00:08:11,250 感觉市场在说 204 00:08:11,250 --> 00:08:12,449 哦大型科技公司 205 00:08:12,449 --> 00:08:13,610 不 206 00:08:13,610 --> 00:08:14,750 没那么好 207 00:08:14,750 --> 00:08:18,029 哦小型亏损公司 208 00:08:18,029 --> 00:08:19,050 非常好 209 00:08:19,050 --> 00:08:20,199 我们喜欢这个 210 00:08:20,199 --> 00:08:21,939 对我来说这没道理 211 00:08:21,939 --> 00:08:24,339 现在完全不继续了 212 00:08:24,339 --> 00:08:24,699 当然 213 00:08:24,699 --> 00:08:25,819 如果你喜欢这类视频 214 00:08:25,819 --> 00:08:26,660 你知道该怎么做 215 00:08:26,660 --> 00:08:27,379 他按了按钮 216 00:08:27,379 --> 00:08:28,430 我们非常感谢 217 00:08:28,430 --> 00:08:29,149 如果有人支持我 218 00:08:29,149 --> 00:08:29,670 甚至更远 219 00:08:29,670 --> 00:08:31,230 Doci在描述区蜷缩 220 00:08:31,230 --> 00:08:32,629 并在评论区冲上前十 221 00:08:32,629 --> 00:08:33,779 目前为止最佳讨论 222 00:08:33,779 --> 00:08:35,299 或前往完整版/斜杠 223 00:08:35,299 --> 00:08:36,120 Goinvestor 224 00:08:36,120 --> 00:08:38,029 非常感谢 225 00:08:38,029 --> 00:08:39,929 所以我们从这里重新开始 226 00:08:39,929 --> 00:08:41,730 关于熵的评论 227 00:08:41,730 --> 00:08:46,299 我们基于半分析的估值基础 228 00:08:46,299 --> 00:08:49,690 半分析他们并不了解 229 00:08:49,690 --> 00:08:50,509 所有这些 230 00:08:50,509 --> 00:08:51,350 人工智能行业 231 00:08:51,350 --> 00:08:53,289 他们只是有份好简讯 232 00:08:53,289 --> 00:08:54,870 他们相当受欢迎 233 00:08:54,870 --> 00:08:56,149 但他们可能出错 234 00:08:56,149 --> 00:08:58,230 而且会经常出错 235 00:08:58,230 --> 00:09:03,950 但他们说我们的估值基础是20倍2027年 236 00:09:03,950 --> 00:09:07,669 3000亿美元终值 237 00:09:07,669 --> 00:09:12,879 2027年每月净新增150亿美元ARR 238 00:09:12,879 --> 00:09:14,220 这将使熵 239 00:09:14,220 --> 00:09:16,539 企业价值约6万亿美元 240 00:09:16,539 --> 00:09:18,899 使其成为全球最大公司 241 00:09:18,899 --> 00:09:20,240 这完全说不通 242 00:09:20,240 --> 00:09:20,659 抱歉 243 00:09:20,659 --> 00:09:22,139 这完全说不通 244 00:09:22,139 --> 00:09:23,210 现在 245 00:09:23,210 --> 00:09:24,830 转向Meta 246 00:09:24,830 --> 00:09:29,509 我认为Meta仍是当前市场最佳故事之一 247 00:09:29,509 --> 00:09:31,610 风险并不大 248 00:09:31,610 --> 00:09:35,620 在我看来Meta应值800美元 249 00:09:35,620 --> 00:09:37,080 或接近1000美元 250 00:09:37,080 --> 00:09:39,139 但这只是我的看法 251 00:09:39,139 --> 00:09:43,419 目前的故事仍然围绕资本支出和良好进展 252 00:09:43,419 --> 00:09:47,019 我想我们接下来要听下一季度或两季度的财报 253 00:09:47,019 --> 00:09:50,789 2027年的资本支出仍将大幅增加 254 00:09:50,789 --> 00:09:52,509 市场可能不会喜欢这个消息 255 00:09:52,509 --> 00:09:55,070 但市场会认可这个决定 256 00:09:56,070 --> 00:09:59,720 如果突然Meta告诉我们，我们实际上已经发布了 257 00:09:59,720 --> 00:10:01,080 新产品和服务 258 00:10:01,080 --> 00:10:04,250 这些产品和服务已被广泛采用 259 00:10:04,250 --> 00:10:06,110 然后市场和分析师会说 260 00:10:06,110 --> 00:10:06,350 哦 261 00:10:06,350 --> 00:10:06,730 你知道的 262 00:10:06,730 --> 00:10:09,850 所有这些资本支出其实都有意义 263 00:10:09,850 --> 00:10:10,529 因为毕竟 264 00:10:10,529 --> 00:10:13,730 这些产品和服务已被大规模采用 265 00:10:13,730 --> 00:10:17,690 这些产品和服务将提升核心AI业务 266 00:10:17,690 --> 00:10:21,629 将为公司带来数十亿美元的额外营收和利润 267 00:10:21,629 --> 00:10:22,730 对公司而言 268 00:10:22,730 --> 00:10:24,710 因此这是合理的 269 00:10:24,710 --> 00:10:27,470 如果我们观察建设进展 270 00:10:27,470 --> 00:10:29,309 我们可以看到加速趋势 271 00:10:29,309 --> 00:10:33,779 所有这些当然基于半导体分析研究 272 00:10:33,779 --> 00:10:38,220 这是每季度计算能力的吉瓦级部署 273 00:10:38,220 --> 00:10:41,460 仅上半年2026年就签订了超过5吉瓦的云和托管合同 274 00:10:41,460 --> 00:10:45,578 2024年以来第三方协议已达近10吉瓦 275 00:10:45,578 --> 00:10:51,210 两个园区正在建设中 276 00:10:51,210 --> 00:10:53,750 各自占2.5吉瓦 277 00:10:53,750 --> 00:10:54,970 我们再次谈论的是Meta 278 00:10:54,970 --> 00:10:57,629 现在 279 00:10:57,629 --> 00:11:00,070 为什么 280 00:11:00,070 --> 00:11:00,350 这不仅是单一举措而是四大战略 281 00:11:00,350 --> 00:11:00,610 我们有路径 282 00:11:00,610 --> 00:11:03,409 第一条外部计算 283 00:11:03,409 --> 00:11:04,330 就像我们看到的SpaceX和Tropic和谷歌 284 00:11:04,330 --> 00:11:05,990 计算协议定价为2.6至4倍 285 00:11:05,990 --> 00:11:10,129 典型五年期云服务费率 286 00:11:10,129 --> 00:11:13,629 类似Meta的200兆瓦协议年收入可达10亿美元 287 00:11:13,629 --> 00:11:16,070 路径二Meta据报正在洽谈私有云访问 288 00:11:16,070 --> 00:11:21,590 效仿亚马逊Bedrock 289 00:11:21,590 --> 00:11:26,529 内部使用令牌即服务 290 00:11:26,529 --> 00:11:28,659 以及可能关联上周报道的垂直应用 291 00:11:28,659 --> 00:11:30,720 正确 292 00:11:30,720 --> 00:11:36,080 Meta云业务报告可能类似路径三 293 00:11:36,080 --> 00:11:36,340 目标是推荐系统实现十倍计算规模 294 00:11:36,340 --> 00:11:41,379 已验证的广告收入生成基础设施 295 00:11:41,379 --> 00:11:45,759 路径三正是当前现实 296 00:11:45,759 --> 00:11:49,559 在第四条路径中 297 00:11:49,559 --> 00:11:53,669 所有容量可立即回拨至前沿训练 298 00:11:53,669 --> 00:11:54,649 建设是AI发展的对冲策略 299 00:11:54,649 --> 00:11:58,330 所有容量均可即时回拨至前沿训练 300 00:11:58,330 --> 00:12:01,009 建设布局是对AI发展的对冲 301 00:12:01,009 --> 00:12:03,688 不要与它赌注 302 00:12:03,688 --> 00:12:06,590 现在这完全合乎逻辑 303 00:12:06,590 --> 00:12:08,929 究竟有什么交易值此代价 304 00:12:08,929 --> 00:12:10,970 SpaceX开创了新市场领域 305 00:12:10,970 --> 00:12:14,570 短期大规模计算以高价提供 306 00:12:14,570 --> 00:12:17,409 SpaceX加谷歌就是这个 307 00:12:17,409 --> 00:12:19,289 其实让我移除自己 308 00:12:19,289 --> 00:12:21,090 所以我看到的就是这个 309 00:12:21,090 --> 00:12:22,440 四十八亿美元 310 00:12:22,440 --> 00:12:24,580 SpaceX熵三十一亿美元 311 00:12:24,580 --> 00:12:26,809 这些都是他们设定的倍数 312 00:12:26,809 --> 00:12:28,470 B300按需计算 313 00:12:28,470 --> 00:12:29,649 二十九亿美元 314 00:12:29,649 --> 00:12:30,649 2.4倍 315 00:12:30,649 --> 00:12:32,929 然后新云AI十二亿美元 316 00:12:32,929 --> 00:12:34,450 这仅是1倍 317 00:12:34,450 --> 00:12:37,909 这基本是每吉瓦年化营收 318 00:12:37,909 --> 00:12:39,769 若Meta达成类似协议 319 00:12:39,769 --> 00:12:43,000 过剩算力成为高毛利收入 320 00:12:43,000 --> 00:12:45,659 市场完全未建模 321 00:12:45,659 --> 00:12:49,100 这引出了最有趣的部分 322 00:12:49,100 --> 00:12:52,259 因为它昨天刚宣布 323 00:12:52,259 --> 00:12:53,879 这就是新图像模型 324 00:12:53,879 --> 00:13:00,690 Muse图像模型在三大主要榜单中排名第二 325 00:13:00,690 --> 00:13:03,370 榜首仍是GPT图像二 326 00:13:03,370 --> 00:13:05,210 我本人也在使用 327 00:13:05,210 --> 00:13:06,210 效果非常出色 328 00:13:06,210 --> 00:13:09,210 而Muse图像现居第二 329 00:13:09,210 --> 00:13:12,460 为何如此重要呢 330 00:13:12,460 --> 00:13:16,019 这很重要因为若关注扎克伯格 331 00:13:16,019 --> 00:13:17,220 你可能看过他的故事 332 00:13:17,220 --> 00:13:21,720 他展示了新图像模型的应用 333 00:13:21,720 --> 00:13:24,419 这并非全新事物 334 00:13:26,330 --> 00:13:28,289 拥有数十亿用户 335 00:13:28,289 --> 00:13:29,450 拥有分发渠道 336 00:13:29,450 --> 00:13:32,470 历史只需一键 337 00:13:32,470 --> 00:13:35,090 即可复制相同操作 338 00:13:35,090 --> 00:13:36,850 但用你的图像 339 00:13:36,850 --> 00:13:38,009 而趋势所在 340 00:13:38,009 --> 00:13:39,009 若某事物流行 341 00:13:39,009 --> 00:13:40,009 作为影响者 342 00:13:40,009 --> 00:13:40,730 大号账号 343 00:13:40,730 --> 00:13:41,590 专业人士 344 00:13:41,590 --> 00:13:43,090 营销公司 345 00:13:43,090 --> 00:13:45,759 任何企业 346 00:13:45,759 --> 00:13:47,440 只需一键 347 00:13:47,440 --> 00:13:48,980 即可完成 348 00:13:48,980 --> 00:13:50,460 发布Instagram动态 349 00:13:50,460 --> 00:13:51,500 新营销视频 350 00:13:51,500 --> 00:13:53,220 这就是其威力 351 00:13:53,220 --> 00:13:59,059 当然，我们一直强调的是，一旦他们拥有实际的基地 352 00:13:59,059 --> 00:14:04,068 这是一个值得开发和升级的良好模型，不断变得更好 353 00:14:04,068 --> 00:14:08,450 随后我们将看到大量产品和服务涌现 354 00:14:08,450 --> 00:14:11,029 然后开始进行分销 355 00:14:11,029 --> 00:14:14,009 只有这样才能实现全面变现 356 00:14:14,009 --> 00:14:15,340 想想看 357 00:14:15,340 --> 00:14:18,019 他们在生态系统中拥有数十亿用户 358 00:14:18,019 --> 00:14:23,210 只需其中很小比例的用户基数 359 00:14:23,210 --> 00:14:25,669 从变现角度来看才有意义 360 00:14:25,669 --> 00:14:27,110 从投资角度来看 361 00:14:27,110 --> 00:14:31,450 但数十亿用户的极小比例 362 00:14:31,450 --> 00:14:33,110 当然就是数十亿 363 00:14:33,110 --> 00:14:35,610 如果不是数百亿美元的营收 364 00:14:35,610 --> 00:14:37,710 高利润率的营收 365 00:14:37,710 --> 00:14:41,129 你拥有优质的视频和图像模型 366 00:14:41,129 --> 00:14:45,210 当然你会看到专业用户和营销人员 367 00:14:45,210 --> 00:14:48,200 中小型企业会使用这些服务 368 00:14:48,200 --> 00:14:48,679 他们确实 369 00:14:48,679 --> 00:14:50,600 当然会额外付费 370 00:14:50,600 --> 00:14:52,519 他们会订阅服务 371 00:14:52,519 --> 00:14:53,500 月度订阅 372 00:14:53,500 --> 00:14:54,320 他们会支付费用 373 00:14:54,320 --> 00:14:56,750 甚至可能为当前使用支付更多 374 00:14:56,750 --> 00:14:58,830 因为他们获得额外价值 375 00:14:58,830 --> 00:15:00,029 这完全合理 376 00:15:00,029 --> 00:15:02,750 这也是我反复强调的 377 00:15:02,750 --> 00:15:07,110 拥有生态系统和分销渠道的一方将胜出 378 00:15:07,110 --> 00:15:08,599 谷歌已经落后了 379 00:15:08,599 --> 00:15:10,438 但看看现在的谷歌 380 00:15:10,438 --> 00:15:11,739 Meta同样落后 381 00:15:11,739 --> 00:15:15,399 但明年Meta会变成什么样呢 382 00:15:15,399 --> 00:15:17,629 你可能会说这不会普及 383 00:15:17,629 --> 00:15:19,070 相信我 384 00:15:19,070 --> 00:15:21,230 Instagram上的数十亿用户 385 00:15:21,230 --> 00:15:25,169 只需要其中一小部分人愿意支付五美元 386 00:15:25,169 --> 00:15:29,529 每月十美元来获取某些云服务 387 00:15:29,529 --> 00:15:34,889 人们愿意支付五到十美元获得更多社交媒体曝光 388 00:15:34,889 --> 00:15:36,250 或许你难以置信 389 00:15:36,250 --> 00:15:42,879 但对于在Meta和社交媒体上投放广告的公司 390 00:15:42,879 --> 00:15:45,580 这无疑是最佳选择 391 00:15:45,580 --> 00:15:47,059 这毫无争议 392 00:15:47,059 --> 00:15:49,710 当然一旦实现这一点 393 00:15:49,710 --> 00:15:51,570 Meta可以展示给我们看 394 00:15:51,570 --> 00:15:53,490 我们已推出这些订阅服务 395 00:15:53,490 --> 00:15:54,649 我们现在拥有 396 00:15:54,649 --> 00:15:55,269 我不知道 397 00:15:55,269 --> 00:15:56,429 十百万订阅用户 398 00:15:56,429 --> 00:15:57,669 两千万订阅用户 399 00:15:57,669 --> 00:15:57,970 无论多少 400 00:15:59,250 --> 00:15:59,450 然后 401 00:15:59,450 --> 00:16:02,549 当然情绪会完全转变 402 00:16:02,549 --> 00:16:04,450 当然从评估角度来看 403 00:16:04,450 --> 00:16:06,250 Meta严重被低估了 404 00:16:06,250 --> 00:16:07,820 我们反复讨论过这个问题 405 00:16:07,820 --> 00:16:10,279 而如往常一样，DCS免费提供 406 00:16:10,279 --> 00:16:10,940 往下看 407 00:16:10,940 --> 00:16:11,840 在置顶评论中 408 00:16:11,840 --> 00:16:14,360 并在描述中提到我 409 00:16:14,360 --> 00:16:16,440 Meta严重被低估了 410 00:16:16,440 --> 00:16:19,960 今天应该值八百美元或更高 411 00:16:19,960 --> 00:16:24,220 我认为一旦我们看到情绪反转 412 00:16:24,220 --> 00:16:27,120 因为Meta向市场展示的是 413 00:16:27,120 --> 00:16:28,399 我们正在实现变现 414 00:16:28,399 --> 00:16:31,559 不仅是核心业务 415 00:16:31,559 --> 00:16:33,980 而是通过订阅模式 416 00:16:33,980 --> 00:16:36,279 或许还有云业务 417 00:16:36,279 --> 00:16:37,419 不管怎样 418 00:16:37,580 --> 00:16:38,460 是的 419 00:16:38,460 --> 00:16:40,860 将会迅速反转 420 00:16:40,860 --> 00:16:43,620 谷歌一年翻倍 421 00:16:43,620 --> 00:16:48,779 为什么Meta不能在未来十二到十八个月内做到同样 422 00:16:48,779 --> 00:16:50,799 我相信他们能做到 423 00:16:50,799 --> 00:16:56,769 如果Entropic在三季度实现十亿美元利润 424 00:16:56,769 --> 00:16:59,110 我认为这对整体非常利好 425 00:16:59,110 --> 00:16:59,970 AI领域 426 00:16:59,970 --> 00:17:01,809 我的问题是 427 00:17:01,809 --> 00:17:06,210 如果他们在季度实现十亿美元利润 428 00:17:06,709 --> 00:17:09,429 使用Entropic的所有公司 429 00:17:09,429 --> 00:17:10,828 是否变得更优秀 430 00:17:10,828 --> 00:17:13,869 公司是否增长更快 431 00:17:13,869 --> 00:17:15,269 如果效率提升 432 00:17:16,269 --> 00:17:18,750 还是这已经是token最大化的顶峰 433 00:17:18,750 --> 00:17:20,750 如果这是token最大化的顶峰 434 00:17:20,750 --> 00:17:25,589 那么四季度和Tropic可能无法达十亿美元利润 435 00:17:25,589 --> 00:17:26,589 这将是问题 436 00:17:26,589 --> 00:17:30,009 这就是我希望这些公司上市的原因 437 00:17:30,009 --> 00:17:34,799 这样我们可以每季度深入观察这些企业 438 00:17:34,799 --> 00:17:36,349 总而言之 439 00:17:36,349 --> 00:17:39,430 我认为市场再次看待这个AI故事 440 00:17:39,430 --> 00:17:41,809 完全错误 441 00:17:41,809 --> 00:17:44,339 我认为我们还有很长的路要走 442 00:17:44,339 --> 00:17:47,099 尤其是像Nebulous这样的公司 443 00:17:47,099 --> 00:17:51,730 我认为Nebulous仍有巨大增长空间 444 00:17:51,730 --> 00:17:55,490 昨天Nebu实际上从六月高点下跌 445 00:17:55,490 --> 00:17:57,819 四个百分点 446 00:17:57,819 --> 00:17:59,539 从六月底 447 00:17:59,539 --> 00:18:01,429 这还不算太久 448 00:18:01,429 --> 00:18:03,108 直到昨日低点 449 00:18:03,108 --> 00:18:04,929 股价下跌约34% 450 00:18:04,929 --> 00:18:09,348 至今仍下跌35%，目前约两百美元每股 451 00:18:09,348 --> 00:18:13,319 我昨天确实买了大约十股额外的股票 452 00:18:13,319 --> 00:18:14,400 不过是啊 453 00:18:14,400 --> 00:18:17,119 对于这些很多名字，我认为即使是核心编织 454 00:18:17,119 --> 00:18:20,710 我认为仍有很长的增长道路要走 455 00:18:20,710 --> 00:18:22,009 他们正在变得 456 00:18:22,009 --> 00:18:24,559 尤其是核心零现在为零 457 00:18:24,559 --> 00:18:26,400 但珊瑚礁与模糊概念有所不同 458 00:18:26,400 --> 00:18:27,200 因为珊瑚礁 459 00:18:27,200 --> 00:18:27,480 嗯 460 00:18:27,480 --> 00:18:28,799 他们向市场做出了承诺 461 00:18:28,799 --> 00:18:29,859 关于利润率方面 462 00:18:29,859 --> 00:18:34,130 所以我想一旦看到珊瑚礁的反转 463 00:18:34,130 --> 00:18:34,809 在利润率层面 464 00:18:34,809 --> 00:18:37,388 我认为这只股票会重返百美元上方 465 00:18:37,388 --> 00:18:38,829 相当轻松 466 00:18:38,829 --> 00:18:40,490 至于半导体板块 467 00:18:40,490 --> 00:18:42,289 无论是英伟达还是AMD 468 00:18:42,289 --> 00:18:43,029 尤其是英伟达 469 00:18:43,029 --> 00:18:44,069 根据这些报告 470 00:18:44,069 --> 00:18:44,230 哦 471 00:18:44,230 --> 00:18:45,150 存在巨大延迟 472 00:18:45,150 --> 00:18:46,710 然后黄仁勋表示没有延迟 473 00:18:46,710 --> 00:18:48,579 路线图依然完整 474 00:18:48,579 --> 00:18:49,980 英伟达目前 475 00:18:49,980 --> 00:18:51,700 从估值角度来看 476 00:18:51,700 --> 00:18:52,339 我认为非常便宜 477 00:18:52,339 --> 00:18:53,670 是的 478 00:18:53,670 --> 00:18:58,150 所有大科技公司都在推出自研芯片等 479 00:18:58,150 --> 00:19:03,130 但他们仍在大量采购英伟达产品 480 00:19:03,130 --> 00:19:04,569 大多数初创公司 481 00:19:04,569 --> 00:19:08,990 如果不是全部都在基于英伟达技术栈开发 482 00:19:08,990 --> 00:19:14,430 因此我不认为英伟达不会保持市场领导地位 483 00:19:14,430 --> 00:19:16,710 但这并不意味着MD不会继续上涨 484 00:19:16,710 --> 00:19:20,920 我仍相信MD将持续快速增长 485 00:19:20,920 --> 00:19:23,319 我们将看到增长加速 486 00:19:23,319 --> 00:19:24,519 今年后半段 487 00:19:24,519 --> 00:19:26,279 从2027年开始 488 00:19:26,279 --> 00:19:29,400 我认为市场存在过度反应，确实如此 489 00:19:29,400 --> 00:19:32,400 我们仍接近市场历史新高 490 00:19:32,400 --> 00:19:32,859 等等 491 00:19:32,859 --> 00:19:33,279 等等 492 00:19:33,279 --> 00:19:34,199 这是事实 493 00:19:34,199 --> 00:19:35,278 这是事实 494 00:19:35,278 --> 00:19:36,838 但如果看英伟达 495 00:19:36,838 --> 00:19:38,199 英伟达正在经历 496 00:19:38,199 --> 00:19:41,278 15%回撤，一度下跌20% 497 00:19:41,278 --> 00:19:46,069 为什么在230美元时那么贵呢 498 00:19:46,069 --> 00:19:47,990 它并非模糊概念 499 00:19:47,990 --> 00:19:48,230 好吧 500 00:19:48,230 --> 00:19:50,390 可以说它大幅上涨 501 00:19:50,390 --> 00:19:52,309 所以小幅回调是合理的 502 00:19:52,309 --> 00:19:54,180 我完全同意你的观点 503 00:19:54,180 --> 00:19:56,740 我们许多公司也具备弹性 504 00:19:56,740 --> 00:19:58,319 甚至是一根轴突 505 00:19:58,319 --> 00:19:58,980 我的意思是 506 00:19:58,980 --> 00:19:59,779 看看这个 507 00:19:59,779 --> 00:20:01,579 这并不正常，所以是的 508 00:20:01,579 --> 00:20:07,380 我不害怕说很多公司在我看来仍然极具吸引力 509 00:20:07,380 --> 00:20:08,559 在当前价格下 510 00:20:08,559 --> 00:20:11,009 我们可能会经历更多下跌交易日 511 00:20:11,009 --> 00:20:12,809 当然会 512 00:20:13,930 --> 00:20:16,829 因为我们仍处于标题驱动的市场 513 00:20:16,829 --> 00:20:17,609 是的 514 00:20:17,609 --> 00:20:20,940 我也可以说有很多股票确实被高估了 515 00:20:20,940 --> 00:20:22,319 价格昂贵 516 00:20:22,319 --> 00:20:25,299 即使你相信这些公司未来价值 517 00:20:25,299 --> 00:20:27,259 会高得多 518 00:20:28,299 --> 00:20:31,210 总之这就是我的看法 519 00:20:31,210 --> 00:20:32,869 在今天的视频中 520 00:20:32,869 --> 00:20:38,190 我将继续讨论Meta直到股价达到800到1000美元 521 00:20:38,190 --> 00:20:42,170 这可能很快就会发生 522 00:20:42,170 --> 00:20:43,730 这就是我今天的视频内容 523 00:20:43,730 --> 00:20:44,930 我们下次再见 524 00:20:44,930 --> 00:20:45,170 再见 525 00:20:45,170 --> 00:20:45,700 再见
评论采集失败
采集完成，该视频无弹幕