WEBVTT - Single Best  Idea with Tom Keene: Frank Lee & Mandeep Singh

0:00:02.520 --> 0:00:07.040
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news.

0:00:14.680 --> 0:00:17.159
<v Speaker 2>Single best idea, which is, if you have two people on,

0:00:17.239 --> 0:00:20.000
<v Speaker 2>you can get a different opinion on all this tech

0:00:20.079 --> 0:00:22.479
<v Speaker 2>and all this mag seven and of course what we

0:00:22.520 --> 0:00:27.920
<v Speaker 2>saw last night in Vidia frankly is an HSBC with

0:00:28.040 --> 0:00:31.320
<v Speaker 2>a wonderful herodage. You're following us. He's got some enthusiasm

0:00:31.640 --> 0:00:36.680
<v Speaker 2>for the industry for the moment at hand. Here on Invidia, Frankly.

0:00:36.800 --> 0:00:39.360
<v Speaker 1>The real map doesn't matter as much as it used to.

0:00:39.680 --> 0:00:41.360
<v Speaker 1>It used to be that Okay and Vidio had the

0:00:41.400 --> 0:00:43.879
<v Speaker 1>best chip, it was leading everybody. But now I think

0:00:43.880 --> 0:00:47.279
<v Speaker 1>what they really need is a new narrative, right, And

0:00:47.600 --> 0:00:51.200
<v Speaker 1>they spent more time talking about server CPUs now because

0:00:51.240 --> 0:00:53.519
<v Speaker 1>this year the big ten, the big trend has been

0:00:53.680 --> 0:00:57.680
<v Speaker 1>agentic AI and suddenly now CPU has gone from a

0:00:57.760 --> 0:01:00.160
<v Speaker 1>runover where nobody cared about to oh, this is the

0:01:00.200 --> 0:01:02.320
<v Speaker 1>next big thing now with the Tentek AI. And so

0:01:02.360 --> 0:01:04.280
<v Speaker 1>I think in video is pivoted now to talk about

0:01:04.280 --> 0:01:05.240
<v Speaker 1>a CPU market.

0:01:05.520 --> 0:01:08.680
<v Speaker 2>So frank leither on the new or the next in Vidia.

0:01:08.720 --> 0:01:11.160
<v Speaker 2>I thought that was fascinating and is an aside. Anna

0:01:11.240 --> 0:01:16.000
<v Speaker 2>Wong was on, and she was for Ana Wong. Gloomy

0:01:16.280 --> 0:01:19.240
<v Speaker 2>about inflation, and she said, okay, you got oil, you

0:01:19.280 --> 0:01:21.960
<v Speaker 2>got the war and all that. Paul sank this week

0:01:22.040 --> 0:01:24.360
<v Speaker 2>saying a gallon of gas it's going to be six dollars.

0:01:24.680 --> 0:01:27.880
<v Speaker 2>I get that. But she moved right on to where

0:01:27.959 --> 0:01:32.320
<v Speaker 2>Frank Lee is on this CPU build out, and Anna

0:01:32.360 --> 0:01:35.800
<v Speaker 2>Wong says, we're going to see a persistent inflation that

0:01:35.880 --> 0:01:38.840
<v Speaker 2>we're all going to face because of the price of

0:01:38.840 --> 0:01:41.600
<v Speaker 2>the chips, the d ram and showing my age. But

0:01:42.160 --> 0:01:44.640
<v Speaker 2>you know, all the different technology stuff is going to

0:01:44.640 --> 0:01:49.200
<v Speaker 2>be an inflationary tendency. That's how our conversations dove till in.

0:01:49.360 --> 0:01:51.680
<v Speaker 2>And so there's Nvidia. I walk in this morning and

0:01:51.720 --> 0:01:54.040
<v Speaker 2>Paul Sweeney's on page two hundred and twenty four of

0:01:54.080 --> 0:01:57.320
<v Speaker 2>two hundred and seventy seven pages of SpaceX. I just

0:01:57.400 --> 0:01:59.960
<v Speaker 2>looked at the photos. I mean, you know, there's beautiful

0:02:00.200 --> 0:02:04.000
<v Speaker 2>photos up front of Mars and the moon and rockets

0:02:04.040 --> 0:02:06.960
<v Speaker 2>going up red Glare and all that. So we did,

0:02:07.000 --> 0:02:10.240
<v Speaker 2>you know, we did a little bit of SPCX the

0:02:10.280 --> 0:02:13.320
<v Speaker 2>red Herring out and we'll see how all that goes.

0:02:13.520 --> 0:02:15.400
<v Speaker 2>Man Deep Sin came in, and of course I want

0:02:15.400 --> 0:02:18.280
<v Speaker 2>to make clear he doesn't do by holed cell and

0:02:18.320 --> 0:02:20.280
<v Speaker 2>he's not going to give you an opinion per se

0:02:20.360 --> 0:02:23.600
<v Speaker 2>on SpaceX, but boy did he have an opinion on

0:02:23.680 --> 0:02:24.400
<v Speaker 2>their accounting.

0:02:24.560 --> 0:02:26.840
<v Speaker 3>I don't get the valuation. I mean, this is a

0:02:26.880 --> 0:02:30.760
<v Speaker 3>company he can come back. Yeah, this is a company

0:02:30.800 --> 0:02:34.720
<v Speaker 3>that's growing thirty percent of a twenty billion dollar revenue braids.

0:02:35.160 --> 0:02:39.080
<v Speaker 3>You have Nvidia that is growing eighty percent of a

0:02:39.080 --> 0:02:42.600
<v Speaker 3>four hundred billion dollar revenue base. Why would you buy

0:02:42.760 --> 0:02:44.639
<v Speaker 3>a Spacexus elons.

0:02:44.840 --> 0:02:46.960
<v Speaker 4>These are the same questions that were being asked of

0:02:47.120 --> 0:02:49.080
<v Speaker 4>Tesla ten years ago.

0:02:49.520 --> 0:02:50.359
<v Speaker 2>Arguably, Yeah, but.

0:02:50.880 --> 0:02:53.360
<v Speaker 3>Look at the growth rates and look at the free

0:02:53.400 --> 0:02:58.040
<v Speaker 3>cash flow for an Nvidia compared to SpaceX. SpaceX is

0:02:58.040 --> 0:02:59.120
<v Speaker 3>not going to show you're going.

0:02:59.080 --> 0:03:01.240
<v Speaker 4>To be one of those guys that's going to be

0:03:01.240 --> 0:03:05.720
<v Speaker 4>focused on valuation. Yeah, okay, you can't cover the stock

0:03:05.760 --> 0:03:07.680
<v Speaker 4>then that's what And I'm not sure we're get we

0:03:07.720 --> 0:03:08.360
<v Speaker 4>gotta find.

0:03:08.200 --> 0:03:12.080
<v Speaker 3>Some trillion dollars for what I mean for the vision.

0:03:12.080 --> 0:03:13.360
<v Speaker 4>Did you read the vision statement?

0:03:13.440 --> 0:03:16.760
<v Speaker 3>Okay, well I read it. There is an Xai portion

0:03:17.200 --> 0:03:19.959
<v Speaker 3>where they're burning capital. And look, they are getting into

0:03:20.000 --> 0:03:24.720
<v Speaker 3>GPU rentals as well. So like SpaceX has this big

0:03:24.800 --> 0:03:27.280
<v Speaker 3>GPU cluster, they're one of the biggest buyers of n

0:03:27.400 --> 0:03:30.280
<v Speaker 3>video chips. Well, guess what, they stop training their model,

0:03:30.360 --> 0:03:33.440
<v Speaker 3>or at least right now, they are renting more of

0:03:33.480 --> 0:03:37.320
<v Speaker 3>their GPU capacity to an entropic as opposed to really

0:03:37.320 --> 0:03:40.080
<v Speaker 3>betting on their own model. So that will generate a

0:03:40.080 --> 0:03:43.080
<v Speaker 3>good fifteen billion dollar revenue run rate just by renting

0:03:43.120 --> 0:03:45.960
<v Speaker 3>the GPUs. But would you pay up you know, eighty

0:03:46.000 --> 0:03:46.920
<v Speaker 3>time sales for that.

0:03:47.360 --> 0:03:50.360
<v Speaker 2>Where's SpaceX going to be in ten years, Where's Man

0:03:50.360 --> 0:03:52.480
<v Speaker 2>Deep Sing going to be in ten years? Where are

0:03:52.480 --> 0:03:54.560
<v Speaker 2>we all going to be in ten years? This is

0:03:54.560 --> 0:03:58.200
<v Speaker 2>what keeps its interesting in economics, finance, investment, and of

0:03:58.240 --> 0:04:03.960
<v Speaker 2>course international relations. On a podcast on on Spotify, on Apple,

0:04:04.280 --> 0:04:12.120
<v Speaker 2>at YouTube podcasts, it's single best idea