1 00:00:02,750 --> 00:00:05,590 Speaker 1: Bloomberg Audio Studios. Podcasts. 2 00:00:05,890 --> 00:00:07,270 Speaker 2: Radio. News. 3 00:00:09,350 --> 00:00:20,840 Speaker 1: Hello and welcome to another episode of the All Thoughts Podcast. 4 00:00:20,980 --> 00:00:21,800 Speaker 1: I'm Traci Allaway. 5 00:00:22,020 --> 00:00:23,020 Speaker 3: And I'm Joe Weisenthal. 6 00:00:23,260 --> 00:00:24,930 Speaker 1: Joe, interesting times in markets. 7 00:00:25,570 --> 00:00:26,030 Speaker 3: You know what? 8 00:00:26,050 --> 00:00:26,130 Speaker 1: What? 9 00:00:27,790 --> 00:00:29,150 Speaker 3: I kind of feel like I lost the plot. 10 00:00:29,390 --> 00:00:29,770 Speaker 1: I know. 11 00:00:29,790 --> 00:00:33,850 Speaker 3: I was like, I had this realization. I don't know 12 00:00:33,900 --> 00:00:36,540 Speaker 3: what anything's going on anymore. We're just talking markets. And 13 00:00:36,580 --> 00:00:38,920 Speaker 3: obviously I lost the plot with life in general. Like 14 00:00:38,960 --> 00:00:41,700 Speaker 3: no one's disputed that. But just as like, what are 15 00:00:41,740 --> 00:00:45,559 Speaker 3: the consensus trades these days? Like what are people actually 16 00:00:45,600 --> 00:00:49,540 Speaker 3: worried about with some of these data center stuff? Like what, Corweave, what, 17 00:00:49,620 --> 00:00:52,379 Speaker 3: et cetera. I have no idea what anyone thinks anymore. 18 00:00:52,440 --> 00:00:54,140 Speaker 3: Like I need a reality check badly. 19 00:00:54,380 --> 00:00:57,910 Speaker 1: So we're recording this on October 28th. And in terms 20 00:00:58,000 --> 00:01:01,240 Speaker 1: of market reality checks, we're actually doing better on this 21 00:01:01,320 --> 00:01:04,399 Speaker 1: one because we often end up recording these right before 22 00:01:04,440 --> 00:01:07,240 Speaker 1: a Fed decision or right before an S-1 comes out 23 00:01:07,340 --> 00:01:11,050 Speaker 1: or right before earnings guidance or something. This time, we've 24 00:01:11,950 --> 00:01:15,690 Speaker 1: timed it pretty well. So we just had the FOMC decision. 25 00:01:15,990 --> 00:01:19,270 Speaker 1: Earlier this week, we had the S-1 filing from Anthropic. 26 00:01:19,810 --> 00:01:23,940 Speaker 1: During the Fed presser, Oracle came out with some additional guidance. 27 00:01:24,200 --> 00:01:26,420 Speaker 1: There's a lot of stuff going on, even though we're 28 00:01:26,440 --> 00:01:27,990 Speaker 1: going to be able to talk about some of it. 29 00:01:28,310 --> 00:01:31,030 Speaker 1: There's more stuff coming up later this week. And so 30 00:01:31,090 --> 00:01:32,690 Speaker 1: I feel like it's a really good moment in time 31 00:01:32,750 --> 00:01:35,660 Speaker 1: to kind of stop and figure out what's going on. 32 00:01:35,670 --> 00:01:40,060 Speaker 3: What's going on? So you mentioned Oracle had some investor 33 00:01:40,080 --> 00:01:43,120 Speaker 3: day data coming out. The stock is up 12%. I know. 34 00:01:43,760 --> 00:01:47,120 Speaker 3: I would say this was a bigger story today than 35 00:01:47,160 --> 00:01:47,700 Speaker 3: the FOMC. 36 00:01:47,980 --> 00:01:48,220 Speaker 1: Yeah. 37 00:01:48,310 --> 00:01:52,500 Speaker 3: Right. Like the questions about Capital expenditures, how much debt 38 00:01:52,620 --> 00:01:56,790 Speaker 3: the big hyperscalers and so forth are going to take on. Clearly, 39 00:01:56,810 --> 00:01:59,450 Speaker 3: I would say in the minds of many participants, much 40 00:01:59,510 --> 00:02:01,830 Speaker 3: bigger questions than like, oh, did they are going to 41 00:02:01,880 --> 00:02:03,000 Speaker 3: hold in October or whatever. 42 00:02:03,440 --> 00:02:08,540 Speaker 1: Not a high bar given Kevin Warsh's ability to deflect. 43 00:02:08,560 --> 00:02:10,200 Speaker 3: I tuned in for a few minutes of that press conference. 44 00:02:10,220 --> 00:02:10,760 Speaker 1: But you're right. 45 00:02:10,780 --> 00:02:11,830 Speaker 3: That guy's good at not saying that. 46 00:02:12,160 --> 00:02:14,370 Speaker 1: All right. Well, we do in fact have the perfect guest. 47 00:02:14,430 --> 00:02:16,410 Speaker 1: We're going to be speaking with someone who we've had 48 00:02:16,470 --> 00:02:18,950 Speaker 1: on the podcast before. Last time we were talking about 49 00:02:19,050 --> 00:02:21,809 Speaker 1: levered ETFs. This time we're going to get a general 50 00:02:21,889 --> 00:02:25,030 Speaker 1: sort of market overview. It's Alex Altman, Global Head of 51 00:02:25,130 --> 00:02:28,070 Speaker 1: Equities Tactical Strategies over at Barclays. So Alex, thank you 52 00:02:28,090 --> 00:02:29,530 Speaker 1: so much for coming back on OddLots. 53 00:02:29,650 --> 00:02:30,970 Speaker 2: Thank you very much for having me again. 54 00:02:31,570 --> 00:02:37,440 Speaker 1: So what's going on? No, I mean, is there like 55 00:02:37,620 --> 00:02:39,980 Speaker 1: a consensus AI outlook at the moment? 56 00:02:40,810 --> 00:02:41,470 Speaker 3: Yeah, there is. 57 00:02:41,720 --> 00:02:46,310 Speaker 2: The consensus AI outlook is that AI is still going to, 58 00:02:46,530 --> 00:02:49,530 Speaker 2: this train is in motion. We're not stopping this train. 59 00:02:49,690 --> 00:02:53,810 Speaker 2: That's still the consensus, to be very clear. And let's 60 00:02:53,850 --> 00:02:55,550 Speaker 2: say if you were to sort of take a sample 61 00:02:55,990 --> 00:03:00,190 Speaker 2: of a bunch of investors across different cohorts, whether it's 62 00:03:00,210 --> 00:03:03,180 Speaker 2: in the US, Europe, Asia, and ask them, gun to 63 00:03:03,220 --> 00:03:05,820 Speaker 2: their head, do you want to still be bullish or 64 00:03:05,860 --> 00:03:09,040 Speaker 2: bearish the AI story? You're going to get an overwhelming 65 00:03:09,160 --> 00:03:10,510 Speaker 2: positive answer. 66 00:03:10,650 --> 00:03:11,150 Speaker 3: Yeah. Right. 67 00:03:11,230 --> 00:03:13,210 Speaker 2: And we can debate like where people are on the 68 00:03:13,230 --> 00:03:17,870 Speaker 2: position scale, but like that's the general view. Right. And 69 00:03:18,510 --> 00:03:21,549 Speaker 2: this is why I think that if you, you know, 70 00:03:21,570 --> 00:03:24,480 Speaker 2: you've probably seen that chart of where we are in 71 00:03:24,720 --> 00:03:27,500 Speaker 2: like the bubble kind of thing. And then as it 72 00:03:27,540 --> 00:03:29,540 Speaker 2: begins to roll over, the first dip down is the 73 00:03:29,560 --> 00:03:30,220 Speaker 2: denial one. 74 00:03:30,260 --> 00:03:30,399 Speaker 1: Yeah. 75 00:03:31,430 --> 00:03:32,910 Speaker 2: I'm not going to say we're at kind of this 76 00:03:32,990 --> 00:03:35,450 Speaker 2: bubbles popped. I know there's commentary out there that actually 77 00:03:35,490 --> 00:03:37,690 Speaker 2: has said that, but I definitely, if you were to 78 00:03:37,730 --> 00:03:40,960 Speaker 2: sort of juxtapose those two narratives, we're in the denial 79 00:03:41,000 --> 00:03:43,880 Speaker 2: phase right now because everyone is still extremely bullish. 80 00:03:43,940 --> 00:03:44,280 Speaker 3: Yeah. 81 00:03:45,600 --> 00:03:47,980 Speaker 2: You got guys saying like, oh, semiconductors, you got Hynix 82 00:03:48,080 --> 00:03:50,440 Speaker 2: or whatever, or some of these Korean memory stocks, they're 83 00:03:50,480 --> 00:03:52,620 Speaker 2: on three times earnings. I'm like, you can look up 84 00:03:52,720 --> 00:03:55,720 Speaker 2: three times earnings on a Bloomberg terminal. That requires no 85 00:03:55,800 --> 00:03:56,280 Speaker 2: work at all. 86 00:03:56,830 --> 00:03:57,030 Speaker 3: Right. 87 00:03:57,090 --> 00:03:58,430 Speaker 2: So everyone is out there. 88 00:03:58,510 --> 00:04:01,309 Speaker 1: Thank you for plugging the utility of the Bloomberg terminal, 89 00:04:01,330 --> 00:04:01,650 Speaker 1: by the way. 90 00:04:01,890 --> 00:04:04,430 Speaker 3: What do you mean there are 3X earnings? 91 00:04:05,490 --> 00:04:06,950 Speaker 2: There are memory stocks that are. 92 00:04:07,270 --> 00:04:10,369 Speaker 3: They're traded with an EPS, not an EPS, a PE 93 00:04:10,430 --> 00:04:14,460 Speaker 3: of three. Correct. Because people there is. And because the 94 00:04:14,510 --> 00:04:17,880 Speaker 3: perception is that we're at such a peak and we 95 00:04:17,940 --> 00:04:21,780 Speaker 3: know how cyclical of an industry it is that when 96 00:04:21,800 --> 00:04:23,720 Speaker 3: the earnings are this good, you're not going to pay 97 00:04:23,779 --> 00:04:26,040 Speaker 3: up for them. Exactly. 98 00:04:26,440 --> 00:04:30,940 Speaker 2: The old mantra is you don't buy cyclical companies, especially 99 00:04:31,050 --> 00:04:35,190 Speaker 2: ones as historically cyclical as semiconductors, when they're at peak earnings. 100 00:04:35,230 --> 00:04:39,150 Speaker 3: I didn't realize they were that low. Okay. So let's 101 00:04:39,450 --> 00:04:41,850 Speaker 3: put some pieces together. So we have the fact that 102 00:04:41,930 --> 00:04:45,179 Speaker 3: the P.E. is actually quite low. People do not expect 103 00:04:45,460 --> 00:04:48,580 Speaker 3: exponential growth from the memory companies in terms of their 104 00:04:48,660 --> 00:04:50,940 Speaker 3: EPS anytime soon as implied by the P.E. 105 00:04:51,020 --> 00:04:51,419 Speaker 2: Ratios. 106 00:04:51,980 --> 00:04:55,200 Speaker 3: What is going on with the neoclouds? Why is it 107 00:04:55,240 --> 00:04:58,219 Speaker 3: that despite, as you say, the consensus is bullish, still 108 00:04:58,260 --> 00:05:01,520 Speaker 3: everyone wants to be long, the AI trade, suddenly there's 109 00:05:01,610 --> 00:05:05,650 Speaker 3: all this anxiety about, you know, people are quoting Corweave 110 00:05:05,690 --> 00:05:06,510 Speaker 3: CDS charts. 111 00:05:06,529 --> 00:05:06,650 Speaker 1: Sure. 112 00:05:06,670 --> 00:05:07,610 Speaker 3: All right. 113 00:05:07,690 --> 00:05:09,710 Speaker 2: So let's take a step back for a second, right? 114 00:05:09,790 --> 00:05:11,790 Speaker 2: Because I want to talk about neoclouds in the context 115 00:05:11,830 --> 00:05:15,850 Speaker 2: of the broader AI ecosystem. So obviously neoclouds are a 116 00:05:15,870 --> 00:05:18,210 Speaker 2: group of businesses that are effectively in some way, shape 117 00:05:18,290 --> 00:05:21,930 Speaker 2: or form mandated to build compute, which they will then 118 00:05:22,029 --> 00:05:26,310 Speaker 2: sell to predetermined customers, some of which predetermined, some of 119 00:05:26,330 --> 00:05:30,880 Speaker 2: them will be speculative. Regardless of what you may believe 120 00:05:30,900 --> 00:05:34,040 Speaker 2: in that business model or not, these are tip of 121 00:05:34,060 --> 00:05:37,400 Speaker 2: the spear. These businesses don't have operating cash flow today 122 00:05:37,480 --> 00:05:40,820 Speaker 2: or minimal operating cash flow today in many cases. And 123 00:05:40,940 --> 00:05:46,330 Speaker 2: so they represent the riskier part of the financing aspect 124 00:05:46,730 --> 00:05:47,770 Speaker 2: of the AI spectrum. 125 00:05:48,100 --> 00:05:48,260 Speaker 1: Right. 126 00:05:48,400 --> 00:05:51,080 Speaker 2: And I think that's, that's an irrefutable fact. You know, 127 00:05:51,100 --> 00:05:53,860 Speaker 2: people can talk about, well, actually they're REITs in disguise 128 00:05:53,920 --> 00:05:56,310 Speaker 2: and actually they're not, they're nowhere near as risky as 129 00:05:56,350 --> 00:05:57,150 Speaker 2: people think they are. 130 00:05:57,670 --> 00:05:57,910 Speaker 1: Sure. 131 00:05:57,970 --> 00:06:01,210 Speaker 2: Let's have that debate. But compared to say the hyperscalers, 132 00:06:01,250 --> 00:06:03,690 Speaker 2: whether it's, you know, the Googles and the Amazons who 133 00:06:03,730 --> 00:06:07,270 Speaker 2: have got huge, I mean, they're cash volcano businesses. They 134 00:06:07,310 --> 00:06:09,990 Speaker 2: are different, different sort of category. And the reason I 135 00:06:10,010 --> 00:06:12,360 Speaker 2: want to make that distinction is because if you think 136 00:06:12,390 --> 00:06:15,320 Speaker 2: of neoclouds as periphery and hyperscalers as core. 137 00:06:15,560 --> 00:06:15,740 Speaker 1: Yeah. 138 00:06:16,220 --> 00:06:18,620 Speaker 2: And I want you to draw a parallel, not directly, 139 00:06:18,760 --> 00:06:20,980 Speaker 2: cause I'm not trying to make some sensationalist statement here, 140 00:06:21,380 --> 00:06:25,000 Speaker 2: but draw a parallel to the GFC. Right. The problem 141 00:06:25,060 --> 00:06:28,760 Speaker 2: doesn't didn't start with GC banks. 142 00:06:29,060 --> 00:06:29,320 Speaker 1: Right. 143 00:06:29,740 --> 00:06:34,599 Speaker 2: It started with subprime vehicles, 12, 18, 24 months beforehand. 144 00:06:34,839 --> 00:06:35,020 Speaker 1: Right. 145 00:06:35,640 --> 00:06:38,840 Speaker 2: And so what's I find most interesting is that there 146 00:06:38,880 --> 00:06:42,980 Speaker 2: is evidently a financing problem within Neo clouds, right? That 147 00:06:43,040 --> 00:06:46,520 Speaker 2: is evident Cypher try to, they did their convert. nuked 148 00:06:46,580 --> 00:06:48,730 Speaker 2: their own share price in doing so because they clearly 149 00:06:48,760 --> 00:06:50,150 Speaker 2: couldn't tap the debt markets directly. 150 00:06:50,170 --> 00:06:51,490 Speaker 3: I don't even know who. Cypher? 151 00:06:51,830 --> 00:06:54,990 Speaker 2: It's another neocloud. Exactly. A company that you've never heard of. 152 00:06:55,589 --> 00:06:58,510 Speaker 2: And yet the reason why CoreWeave bonds sold off was 153 00:06:58,570 --> 00:07:00,850 Speaker 2: not because of CoreWeave. They didn't say anything that week. 154 00:07:01,350 --> 00:07:06,690 Speaker 2: It was because Cypher, another neocloud, tried to tap financing. 155 00:07:07,170 --> 00:07:10,090 Speaker 2: And in doing so, materially hurt their own share price. 156 00:07:10,750 --> 00:07:13,230 Speaker 2: And then all of a sudden, a bunch of investors, 157 00:07:13,830 --> 00:07:15,410 Speaker 2: quick off the mark, put two and two together. 158 00:07:16,010 --> 00:07:16,210 Speaker 1: Right. 159 00:07:16,570 --> 00:07:18,630 Speaker 2: And then it spilled over, but like, quite frankly, if 160 00:07:18,670 --> 00:07:21,150 Speaker 2: you look at the data center bonds in isolation, whether 161 00:07:21,170 --> 00:07:22,450 Speaker 2: like the beignets of this. 162 00:07:22,370 --> 00:07:25,110 Speaker 3: World, you know, the meta one, exactly. 163 00:07:25,170 --> 00:07:25,350 Speaker 1: Yeah. 164 00:07:25,590 --> 00:07:29,210 Speaker 2: These had already been widening. For months prior to that. 165 00:07:29,550 --> 00:07:29,710 Speaker 1: Right. 166 00:07:29,750 --> 00:07:32,650 Speaker 2: And we'd seen a creep in IG hyperscale is one of. 167 00:07:32,650 --> 00:07:36,130 Speaker 3: These ones that pivoted from Bitcoin mining to getting into AI. 168 00:07:36,190 --> 00:07:37,320 Speaker 1: Yeah. Yeah. All right. 169 00:07:37,360 --> 00:07:40,160 Speaker 2: So I guess the point I'm trying to make is, 170 00:07:40,240 --> 00:07:44,050 Speaker 2: is that there's ripple effects in financing that start in. 171 00:07:44,760 --> 00:07:47,120 Speaker 2: in the periphery, and we're seeing that. So what I 172 00:07:47,140 --> 00:07:48,700 Speaker 2: find amazing, back to the original point in terms of 173 00:07:48,800 --> 00:07:50,800 Speaker 2: you asked what's going on, are people bullish or bearish, 174 00:07:51,300 --> 00:07:54,260 Speaker 2: and we literally just conducted a survey with a group 175 00:07:54,290 --> 00:07:55,030 Speaker 2: of investors. 176 00:07:55,390 --> 00:07:55,750 Speaker 1: Fantastic. 177 00:07:56,070 --> 00:07:58,950 Speaker 2: Yeah. And the only, I mean, you could make this argument, 178 00:07:58,970 --> 00:08:01,550 Speaker 2: this is a high number, but from the responses that 179 00:08:01,570 --> 00:08:04,990 Speaker 2: we've got so far, around about 60% of investors, we 180 00:08:05,010 --> 00:08:07,130 Speaker 2: asked them a question, do you think financing is going 181 00:08:07,150 --> 00:08:08,450 Speaker 2: to be a problem? AI financing is going to be 182 00:08:08,470 --> 00:08:10,230 Speaker 2: a problem for the foreseeable future. 183 00:08:10,690 --> 00:08:11,280 Speaker 3: Only 60% said yes. 184 00:08:12,970 --> 00:08:16,130 Speaker 2: By definition, what's happened over the past couple of weeks 185 00:08:16,290 --> 00:08:20,910 Speaker 2: means the answer should be 100%, right? And yet, let's 186 00:08:20,950 --> 00:08:23,190 Speaker 2: just call it just over half think or just under 187 00:08:23,230 --> 00:08:25,610 Speaker 2: half think there is no financing issue, but there is. 188 00:08:26,300 --> 00:08:30,140 Speaker 2: So what's happening in AI to me is evidence of 189 00:08:30,180 --> 00:08:33,140 Speaker 2: like almost a denial of a reality that's already here. 190 00:08:33,590 --> 00:08:37,069 Speaker 1: To be fair, one of the reasons CoreWeave started selling 191 00:08:37,130 --> 00:08:39,410 Speaker 1: off was because Brandon McBee gave an interview to the 192 00:08:39,510 --> 00:08:43,920 Speaker 1: FT where he was talking literally about unlimited amounts of borrowing, 193 00:08:43,960 --> 00:08:47,760 Speaker 1: which I think ties into what you're discussing when it 194 00:08:47,800 --> 00:08:53,559 Speaker 1: comes to capital concerns. So on the financing side, I mean, 195 00:08:53,660 --> 00:08:59,020 Speaker 1: we've seen some differentiation in terms of lending money to 196 00:08:59,260 --> 00:09:04,079 Speaker 1: different AI entities out there. Is the overarching point that 197 00:09:04,140 --> 00:09:06,500 Speaker 1: we've been in a period where it seemed like there 198 00:09:06,540 --> 00:09:09,200 Speaker 1: was in fact unlimited money and people could tap the 199 00:09:09,340 --> 00:09:11,800 Speaker 1: IG markets or the loan markets for as much as 200 00:09:11,840 --> 00:09:14,160 Speaker 1: they needed. And now we're moving to a period where 201 00:09:14,420 --> 00:09:16,620 Speaker 1: things are more selective and not everyone is going to 202 00:09:16,660 --> 00:09:19,280 Speaker 1: be able to get the type of funding that they 203 00:09:19,320 --> 00:09:20,540 Speaker 1: were tapping before. 204 00:09:21,179 --> 00:09:22,280 Speaker 3: That is exactly it. 205 00:09:22,760 --> 00:09:27,020 Speaker 2: Right. Because, you know, part of the job that my 206 00:09:27,040 --> 00:09:29,089 Speaker 2: team does is we travel the world. We talk to 207 00:09:29,110 --> 00:09:31,880 Speaker 2: a lot of different investors and, We ask questions, we 208 00:09:31,940 --> 00:09:35,560 Speaker 2: collate the kind of feedback that informs our decisions. And 209 00:09:35,600 --> 00:09:39,480 Speaker 2: there was this almost universal view that, and you guys 210 00:09:39,500 --> 00:09:42,040 Speaker 2: would have heard it as well, AI was immune to 211 00:09:42,080 --> 00:09:47,970 Speaker 2: high interest rates. Evidently is not true, right? And so 212 00:09:48,000 --> 00:09:51,090 Speaker 2: there is a spectrum. And even, I think that's where 213 00:09:51,110 --> 00:09:54,610 Speaker 2: I think the meta results were pretty instructive in the 14C, 214 00:09:55,290 --> 00:09:58,809 Speaker 2: in the sense that they highlighted the fact that financing 215 00:09:58,950 --> 00:10:02,990 Speaker 2: does constrain, financing does have limitations. And we wrote a 216 00:10:03,110 --> 00:10:04,370 Speaker 2: load of, we did a load of work on this 217 00:10:04,390 --> 00:10:06,929 Speaker 2: in the summer where we were looking, not specifically at AI, 218 00:10:06,980 --> 00:10:10,120 Speaker 2: we were just looking at the holistic net issuance profile 219 00:10:10,130 --> 00:10:12,760 Speaker 2: of the market. And maybe you guys saw this. And 220 00:10:13,120 --> 00:10:16,020 Speaker 2: we looked at just the entire corporate market, so IG 221 00:10:16,059 --> 00:10:19,400 Speaker 2: high yield. We looked at treasuries and we looked at, equity. 222 00:10:19,940 --> 00:10:22,860 Speaker 2: And we went back to about 1995, collated all the data, 223 00:10:23,520 --> 00:10:25,900 Speaker 2: and looked at each market, the amount of net issuance 224 00:10:25,950 --> 00:10:28,429 Speaker 2: in any given year as a percentage of each market's 225 00:10:28,970 --> 00:10:31,949 Speaker 2: market cap in that given year. And then we plotted 226 00:10:31,970 --> 00:10:34,590 Speaker 2: it over time. And what we found was there was, unsurprisingly, 227 00:10:34,630 --> 00:10:37,910 Speaker 2: a very strong relationship. When interest rates went down, overall 228 00:10:37,950 --> 00:10:40,290 Speaker 2: issuance went up, and vice versa, as you'd expect. 229 00:10:40,700 --> 00:10:42,240 Speaker 3: The laws of the economy still exist. 230 00:10:42,770 --> 00:10:45,390 Speaker 2: Well, apparently not so much in 2025 and 2026. 231 00:10:45,390 --> 00:10:47,410 Speaker 3: Yeah, here's the punchline. 232 00:10:47,460 --> 00:10:50,680 Speaker 2: Yeah, this is the... Yeah, exactly. This is the punchline. 233 00:10:50,679 --> 00:10:54,990 Speaker 2: 2026 has completely shattered what you'd seen in the norms. 234 00:10:55,010 --> 00:10:57,540 Speaker 2: I think Tracy used You looked at that chart. Yeah. 235 00:10:57,559 --> 00:10:59,959 Speaker 1: So you had, you had this great chart plotting the 236 00:11:00,000 --> 00:11:03,480 Speaker 1: relationship between the two year yield and net issuance. And 237 00:11:03,520 --> 00:11:07,620 Speaker 1: you can see the relationship really breaks down in 2026. 238 00:11:07,620 --> 00:11:11,959 Speaker 1: So we are issuing a lot more. We're selling a 239 00:11:11,980 --> 00:11:15,420 Speaker 1: lot more bonds, a lot more debt related to benchmark 240 00:11:15,480 --> 00:11:19,240 Speaker 1: rates than pretty much like ever in history. It's, it's 241 00:11:19,280 --> 00:11:20,020 Speaker 1: kind of biblical. 242 00:11:20,420 --> 00:11:22,660 Speaker 2: It's certainly, we have data back to the nineties. So, 243 00:11:22,679 --> 00:11:25,210 Speaker 2: you know, if you compare it to say.com, And the 244 00:11:25,230 --> 00:11:27,010 Speaker 2: interesting thing about dot-com is, of course, there was a 245 00:11:27,030 --> 00:11:30,000 Speaker 2: lot of equity issuance and interest rates were high, but 246 00:11:30,120 --> 00:11:32,059 Speaker 2: the US government was basically running a surplus and all 247 00:11:32,080 --> 00:11:34,840 Speaker 2: this kind of stuff. So today is very unique. And 248 00:11:34,860 --> 00:11:36,559 Speaker 2: the reason why I bring it back to the sort 249 00:11:36,660 --> 00:11:40,100 Speaker 2: of AI financing is because there is this competition for capital. 250 00:11:40,340 --> 00:11:43,120 Speaker 2: It is real. AI is not issuing debt in a vacuum. 251 00:11:43,160 --> 00:11:46,260 Speaker 2: There are other verticals that need to tap financing too. 252 00:11:46,860 --> 00:11:49,600 Speaker 2: And there is a finite amount, right? So IG issuance 253 00:11:50,000 --> 00:11:52,719 Speaker 2: and a high-yield market this year on a net basis, 254 00:11:52,780 --> 00:11:58,020 Speaker 2: we forecast to be $ 1. 1 trillion, right? And we just 255 00:11:58,070 --> 00:12:01,610 Speaker 2: throw around trillions as though it's just this casual number 256 00:12:01,630 --> 00:12:04,900 Speaker 2: that everyone's got, but there is limitations, right? 257 00:12:21,690 --> 00:12:26,510 Speaker 3: Reconcile this. If there is this perception that people are 258 00:12:26,550 --> 00:12:31,089 Speaker 3: still bullish, there is a, the train is still moving 259 00:12:31,130 --> 00:12:35,640 Speaker 3: at full speed. Why should capital be hesitant? Like, okay. 260 00:12:35,679 --> 00:12:38,560 Speaker 3: So it's like, you can, you can see on the numbers, 261 00:12:38,640 --> 00:12:42,179 Speaker 3: like the core weaves and the irons and the cipher 262 00:12:42,260 --> 00:12:43,660 Speaker 3: and like a bunch of other ones that I had 263 00:12:43,740 --> 00:12:47,280 Speaker 3: never heard of. Like, yes, clearly there's the capital markets 264 00:12:47,340 --> 00:12:50,780 Speaker 3: are telling something, telling them something. But why? If the 265 00:12:50,840 --> 00:12:55,100 Speaker 3: view is that AI is still generally going straight up, 266 00:12:55,559 --> 00:12:57,979 Speaker 3: what is the fear about these companies that are at 267 00:12:58,020 --> 00:13:02,500 Speaker 3: the tip of the sphere? Wouldn't the bull say, look, 268 00:13:02,850 --> 00:13:06,790 Speaker 3: someone is going to pay for these GPUs. The demand 269 00:13:06,870 --> 00:13:09,689 Speaker 3: is going up. If they buy or rent or whatever, 270 00:13:09,730 --> 00:13:12,150 Speaker 3: if they have electricity, if they have GPUs, if they 271 00:13:12,170 --> 00:13:15,310 Speaker 3: bought them, whatever, they're not going to zero. They're going 272 00:13:15,330 --> 00:13:16,750 Speaker 3: to be worth a lot of money somewhere or nothing. 273 00:13:17,370 --> 00:13:17,550 Speaker 1: Yeah. 274 00:13:17,730 --> 00:13:20,930 Speaker 2: So that, that is the bull argument is there's nothing 275 00:13:20,970 --> 00:13:22,530 Speaker 2: wrong with it in terms of build it and they 276 00:13:22,550 --> 00:13:25,630 Speaker 2: will come. I would say that, that the challenge will 277 00:13:25,670 --> 00:13:28,310 Speaker 2: be to the early two things to the earlier point, 278 00:13:28,790 --> 00:13:30,949 Speaker 2: tip of the spear. So when you start to see 279 00:13:31,030 --> 00:13:33,650 Speaker 2: cracks or canaries in coal mines or however you want 280 00:13:33,670 --> 00:13:36,610 Speaker 2: to characterize it, you should probably pay attention. Right. 281 00:13:36,710 --> 00:13:38,850 Speaker 3: Um, and, and. 282 00:13:38,809 --> 00:13:41,250 Speaker 2: The second point is, is that, thank you. Uh, the 283 00:13:41,270 --> 00:13:45,380 Speaker 2: second point is we, we need to think about revenues 284 00:13:45,440 --> 00:13:47,329 Speaker 2: at the end of the day. And this is something 285 00:13:47,350 --> 00:13:48,730 Speaker 2: which we addressed at the end of August in a 286 00:13:48,809 --> 00:13:52,330 Speaker 2: note called placing your AI chips more carefully. And the 287 00:13:52,429 --> 00:13:56,550 Speaker 2: idea was, is that we basically decomposed. It was around 288 00:13:56,580 --> 00:14:00,360 Speaker 2: the time of the SpaceX TAM. If you remember, that 289 00:14:00,380 --> 00:14:01,960 Speaker 2: was around about $ 30 trillion. 290 00:14:02,059 --> 00:14:07,000 Speaker 1: It was a total addressable market of like the entire universe, 291 00:14:07,120 --> 00:14:08,910 Speaker 1: not just earth, something like that. 292 00:14:08,950 --> 00:14:11,929 Speaker 2: Yeah. $ 30 trillion, which is obviously around about the GDP 293 00:14:11,970 --> 00:14:14,830 Speaker 2: of the United States. And so we went, again, to 294 00:14:14,850 --> 00:14:17,690 Speaker 2: the point about people throwing around trillions. Fun fact for you. 295 00:14:18,570 --> 00:14:21,310 Speaker 2: Since 1990, I can't actually take credit for this stat, 296 00:14:21,350 --> 00:14:23,990 Speaker 2: by the way. This actually belongs to Cameron Crease at, 297 00:14:24,150 --> 00:14:24,890 Speaker 2: who also works here? 298 00:14:25,010 --> 00:14:26,030 Speaker 3: Our colleague, yeah. 299 00:14:26,150 --> 00:14:31,280 Speaker 2: And so give him the necessary plug on this. Looking 300 00:14:31,360 --> 00:14:36,120 Speaker 2: at the cumulative revenue of the two most successful tech 301 00:14:36,160 --> 00:14:39,920 Speaker 2: companies in modern history, it was Microsoft and Apple, and 302 00:14:40,240 --> 00:14:44,720 Speaker 2: it was the cumulative TAM of those two was, I think, inflation-adjusted, 303 00:14:44,780 --> 00:14:50,650 Speaker 2: not even $ 10 trillion since 1990. But SpaceX's addressable is 304 00:14:50,930 --> 00:14:55,830 Speaker 2: talking three times that number. So that kind of prompts us, 305 00:14:55,870 --> 00:14:59,410 Speaker 2: where's all this money going to come from? So punchline again, 306 00:14:59,410 --> 00:15:03,600 Speaker 2: $ 17 trillion of S & P top line revenue. And 307 00:15:03,620 --> 00:15:06,600 Speaker 2: they've got around about $ 11 or $ 12 trillion of expenses. 308 00:15:07,340 --> 00:15:09,840 Speaker 2: And so that means that the net profit approximately of 309 00:15:09,900 --> 00:15:13,530 Speaker 2: the S & P is somewhere between five and five 310 00:15:13,550 --> 00:15:14,690 Speaker 2: and a half trillion. 311 00:15:14,370 --> 00:15:14,950 Speaker 3: Dollars, right? 312 00:15:15,370 --> 00:15:17,370 Speaker 2: So it's like, all right, we can make some bold 313 00:15:17,430 --> 00:15:19,250 Speaker 2: claims about the amount of money that's going to get 314 00:15:19,350 --> 00:15:22,650 Speaker 2: allocated to AI expenditures and all the rest of it. 315 00:15:23,310 --> 00:15:25,510 Speaker 2: But last time I checked, like we all look at 316 00:15:25,550 --> 00:15:28,250 Speaker 2: the ramp data or the open router data, and we 317 00:15:28,270 --> 00:15:31,120 Speaker 2: can clearly see that that token expenditure has kind of 318 00:15:31,160 --> 00:15:34,260 Speaker 2: begun to level off a bit, right? At least near term. 319 00:15:34,280 --> 00:15:36,600 Speaker 2: So it's like, all right, our company is going to 320 00:15:36,640 --> 00:15:40,380 Speaker 2: spend this much money. And that's where the other point 321 00:15:40,420 --> 00:15:41,520 Speaker 2: to answer your question, Joe, is. 322 00:15:41,810 --> 00:15:45,770 Speaker 3: Well, you know, your company Barclays is, I know, managing 323 00:15:45,790 --> 00:15:49,000 Speaker 3: the anthropic IPO. So, or as, you know, involved in this, 324 00:15:49,040 --> 00:15:52,280 Speaker 3: we're not going to ask any anthropic questions to you, 325 00:15:52,660 --> 00:15:54,060 Speaker 3: but I will say to Tracy. 326 00:15:54,220 --> 00:15:54,540 Speaker 1: Okay. 327 00:15:54,940 --> 00:15:57,140 Speaker 3: No, I won't make you comment. You know, we understand 328 00:15:57,160 --> 00:16:00,910 Speaker 3: the rules here, but I will say to Tracy that 329 00:16:01,490 --> 00:16:05,109 Speaker 3: anthropic numbers on the top line look great. And I 330 00:16:05,130 --> 00:16:07,570 Speaker 3: do not think that nobody, and I don't think we 331 00:16:07,590 --> 00:16:12,119 Speaker 3: learned that much about open route, open router data. because 332 00:16:12,180 --> 00:16:14,860 Speaker 3: that captures a specific slice of people who are usually 333 00:16:14,900 --> 00:16:18,460 Speaker 3: looking for cheap Chinese models, which they're doing well, but 334 00:16:18,500 --> 00:16:21,390 Speaker 3: the top line numbers for Anthropic look unbelievably good. That's 335 00:16:21,410 --> 00:16:21,970 Speaker 3: the interesting. 336 00:16:22,410 --> 00:16:26,470 Speaker 1: Thank you. There's earnings progress there in the S1. You 337 00:16:26,490 --> 00:16:31,840 Speaker 1: can see the earnings progression. However, billions in earnings versus 338 00:16:31,860 --> 00:16:35,560 Speaker 1: valuations at trillions, I think that is the gap that 339 00:16:35,600 --> 00:16:36,400 Speaker 1: people are debating. 340 00:16:36,440 --> 00:16:39,710 Speaker 3: I'm just pointing out that There's lots of charts. There's 341 00:16:39,790 --> 00:16:42,340 Speaker 3: ramp data that's floating around. Lots of people look at 342 00:16:42,360 --> 00:16:44,860 Speaker 3: the ramp data and lots of people look at the 343 00:16:44,880 --> 00:16:49,580 Speaker 3: statistics on OpenRouter. And they start to say, oh, the 344 00:16:49,600 --> 00:16:52,340 Speaker 3: American Frontier Labs, I guess, are they in trouble? And 345 00:16:52,360 --> 00:16:54,210 Speaker 3: they might be. But I'm just saying, if you look 346 00:16:54,230 --> 00:16:56,970 Speaker 3: at the data that's filed with the SEC, as of 347 00:16:57,010 --> 00:16:59,430 Speaker 3: pretty recently, at least one of the companies, the one 348 00:16:59,470 --> 00:17:02,410 Speaker 3: that is filed in S1, that number is going straight 349 00:17:02,450 --> 00:17:02,810 Speaker 3: up still. 350 00:17:03,080 --> 00:17:06,840 Speaker 1: Why don't we talk about the frontier models more generally? So, 351 00:17:07,140 --> 00:17:09,020 Speaker 1: you know, the argument has been that they can come 352 00:17:09,080 --> 00:17:12,040 Speaker 1: up with a better model, a better agent, a better technology, 353 00:17:12,180 --> 00:17:14,220 Speaker 1: and that people are ultimately going to pay for it. 354 00:17:14,280 --> 00:17:17,000 Speaker 1: And the persistent, I guess, doubt in the market has been, well, 355 00:17:17,100 --> 00:17:19,359 Speaker 1: you know, open source models are kind of catching up 356 00:17:19,400 --> 00:17:23,540 Speaker 1: very quickly. They're worse on their benchmark test, but not 357 00:17:23,619 --> 00:17:26,820 Speaker 1: by that much anymore. And they're a lot cheaper to operate. 358 00:17:27,200 --> 00:17:29,869 Speaker 1: What are people saying about the open source versus frontier debate? 359 00:17:31,400 --> 00:17:35,159 Speaker 2: I would say here, this is weirdly, is a bit 360 00:17:35,180 --> 00:17:39,440 Speaker 2: of a contradiction versus that bullish narrative. Because the data 361 00:17:39,460 --> 00:17:41,480 Speaker 2: is the data you can evidently see on a whole 362 00:17:41,520 --> 00:17:46,139 Speaker 2: bunch of different sources that open weight has kind of 363 00:17:46,369 --> 00:17:49,030 Speaker 2: gone from a 20% market share to anything up to 364 00:17:49,070 --> 00:17:53,250 Speaker 2: about an 80% market share. And then you'll also see 365 00:17:53,310 --> 00:17:58,270 Speaker 2: that the token cost is a fraction, one one hundredth. 366 00:17:58,520 --> 00:18:01,619 Speaker 3: Versus- Incredible token deflation that's happened every year. 367 00:18:01,700 --> 00:18:04,840 Speaker 2: It might be one of the fastest deflating assets in history. 368 00:18:04,900 --> 00:18:06,580 Speaker 2: It makes those NFTs back in 2021 too. 369 00:18:06,600 --> 00:18:10,820 Speaker 3: But GPU rental rates have been going up. They have. 370 00:18:11,280 --> 00:18:14,379 Speaker 2: Yes, absolutely. And so there is clearly a two-way debate. 371 00:18:14,480 --> 00:18:16,540 Speaker 2: And I think it again goes back to the evolution 372 00:18:16,560 --> 00:18:19,320 Speaker 2: of this broader AI narrative. It's not just necessarily a 373 00:18:19,400 --> 00:18:21,580 Speaker 2: one-way train, rising tide lifts all boats. There's going to 374 00:18:21,600 --> 00:18:25,580 Speaker 2: be degrees of differentiation. But you asked about open weight models. 375 00:18:25,619 --> 00:18:27,800 Speaker 2: And I think where one of the leading edge questions 376 00:18:27,840 --> 00:18:32,140 Speaker 2: right now is not even just open weight, but combine 377 00:18:32,160 --> 00:18:37,119 Speaker 2: that with locally run open weight. And this is something 378 00:18:37,160 --> 00:18:42,230 Speaker 2: which I like to call being long or short human creativity. 379 00:18:42,770 --> 00:18:43,010 Speaker 1: Right. 380 00:18:43,369 --> 00:18:47,810 Speaker 2: In so far that if, when the iPhone came out 381 00:18:47,830 --> 00:18:50,890 Speaker 2: in 2007, stay with me here, uh, that if I had, 382 00:18:51,250 --> 00:18:53,800 Speaker 2: if someone from the future had said 20 years from now, 383 00:18:53,910 --> 00:18:57,080 Speaker 2: you're going to have whatever the orders of magnitude compute 384 00:18:57,119 --> 00:18:59,979 Speaker 2: differential And the battery is going to be the same size. 385 00:19:00,200 --> 00:19:04,000 Speaker 2: That 2007 person would have said impossible. You'd need an 386 00:19:04,020 --> 00:19:07,389 Speaker 2: iPhone the size of this table, right? So clearly we 387 00:19:07,430 --> 00:19:10,490 Speaker 2: got there amongst many other things. And so the idea 388 00:19:10,570 --> 00:19:15,030 Speaker 2: that we necessarily need to ping a bunch of hoppers 389 00:19:15,450 --> 00:19:19,160 Speaker 2: or Vera Rubins in an Ohio data center just so 390 00:19:19,200 --> 00:19:21,660 Speaker 2: that we can sort of get our. 391 00:19:21,600 --> 00:19:23,139 Speaker 1: Agent to- Manage my email for me. 392 00:19:23,180 --> 00:19:26,980 Speaker 2: Yeah, or to send an angry response letter to the hotel. 393 00:19:27,000 --> 00:19:28,080 Speaker 3: You're going to be able to do that with the 394 00:19:28,119 --> 00:19:29,420 Speaker 3: chip on your phone by next year. 395 00:19:30,020 --> 00:19:32,419 Speaker 2: Right, if not already. So we've all kind of got 396 00:19:32,460 --> 00:19:36,380 Speaker 2: supercomputers in our pocket. And the ability to run, say, 397 00:19:36,500 --> 00:19:41,550 Speaker 2: multi-billion parameter models locally using whether an open-weight model or 398 00:19:41,590 --> 00:19:45,650 Speaker 2: a closed-weight model, but that's provided by NVIDIA's Nematron or 399 00:19:46,050 --> 00:19:49,250 Speaker 2: whatever Meta's putting out, which Meta's already put out a- 400 00:19:50,080 --> 00:19:52,800 Speaker 2: multi-billion parameter model that's run locally on your laptop. 401 00:19:52,880 --> 00:19:54,080 Speaker 3: It's crazy. Right. 402 00:19:54,820 --> 00:19:58,659 Speaker 2: But it's not because that technology, right? And so I 403 00:19:58,700 --> 00:20:00,540 Speaker 2: think that's the other point of the debate is it's 404 00:20:00,580 --> 00:20:02,840 Speaker 2: not even so much about open weight. It's about the 405 00:20:02,920 --> 00:20:06,540 Speaker 2: fact that you're moving along this efficiency curve And that 406 00:20:06,640 --> 00:20:09,700 Speaker 2: was also bringing in the revenue question as well, which 407 00:20:10,040 --> 00:20:13,270 Speaker 2: ties in with those frontier models as well. Frontier models, 408 00:20:13,830 --> 00:20:16,229 Speaker 2: we're not going to be producing, you and I, with 409 00:20:16,250 --> 00:20:18,570 Speaker 2: the greatest respect to everyone in this room, we're probably 410 00:20:18,630 --> 00:20:21,350 Speaker 2: not going to be discovering the new biotechnological drugs. 411 00:20:38,010 --> 00:20:39,850 Speaker 3: I was actually, I was talking to someone when I 412 00:20:39,869 --> 00:20:42,449 Speaker 3: was out in San Francisco recently, and I actually want 413 00:20:42,470 --> 00:20:45,030 Speaker 3: to ask you about East and West Coast views on 414 00:20:45,050 --> 00:20:47,409 Speaker 3: the technology, maybe. But I was talking to someone in 415 00:20:47,850 --> 00:20:50,869 Speaker 3: San Francisco who had this term, he called it the 416 00:20:50,930 --> 00:20:54,530 Speaker 3: frontier sink. And it's like, what are the industries that 417 00:20:54,770 --> 00:20:59,710 Speaker 3: actually need PhD level and beyond technology? And there's a 418 00:20:59,730 --> 00:21:02,020 Speaker 3: few that we could come up with pretty easily. One 419 00:21:02,040 --> 00:21:04,880 Speaker 3: of them is probably biotech, right? You can probably never 420 00:21:04,940 --> 00:21:08,359 Speaker 3: have too much intelligence when it comes to medicine. Another 421 00:21:08,460 --> 00:21:11,860 Speaker 3: one actually is probably semiconductor design itself. You'll probably never 422 00:21:11,940 --> 00:21:16,260 Speaker 3: be too smart, et cetera, for that. Maybe another one 423 00:21:16,280 --> 00:21:19,119 Speaker 3: of them has to do with quant trading where you 424 00:21:19,140 --> 00:21:21,550 Speaker 3: can never be too smart. I'm not sure. Who knows? 425 00:21:22,210 --> 00:21:25,610 Speaker 3: But it actually does get to be hard to think about. 426 00:21:25,970 --> 00:21:29,330 Speaker 3: There might just be a lot of industry for which 427 00:21:30,010 --> 00:21:32,890 Speaker 3: there is a certain level of intelligence that is enough. 428 00:21:34,050 --> 00:21:36,530 Speaker 2: I mean, for the moment, again, Do you remember when the, 429 00:21:37,080 --> 00:21:39,720 Speaker 2: I like going back to the iPhone as an analogy, 430 00:21:39,780 --> 00:21:42,879 Speaker 2: because it really was a technological revolution at the time. 431 00:21:43,270 --> 00:21:45,350 Speaker 2: And do you remember when we thought like peak iPhone 432 00:21:45,470 --> 00:21:48,090 Speaker 2: in 07, 08 was when you had that app, which 433 00:21:48,190 --> 00:21:51,030 Speaker 2: had like the pint of beer in it. 434 00:21:51,070 --> 00:21:52,890 Speaker 1: And you could drink it. 435 00:21:52,970 --> 00:21:55,340 Speaker 3: I didn't, but no, everyone is like, Oh, this is 436 00:21:55,380 --> 00:21:57,580 Speaker 3: so cool. You can pretend. I know exactly. 437 00:21:57,600 --> 00:22:00,899 Speaker 1: Joe, you're allowed to take joy in small things. No, no. 438 00:22:01,890 --> 00:22:04,010 Speaker 3: This looks like whimsy to me. I don't get this. Oh, yeah. 439 00:22:04,050 --> 00:22:04,690 Speaker 1: Joe hates whimsy. 440 00:22:04,730 --> 00:22:07,149 Speaker 3: I hate fun. But keep going. But I remember how, yo, 441 00:22:07,350 --> 00:22:09,550 Speaker 3: why is everyone going crazy about the fact that you could, like, 442 00:22:10,310 --> 00:22:11,630 Speaker 3: this is a very good point. Keep going. 443 00:22:11,750 --> 00:22:14,730 Speaker 2: And so I just don't think we know what the 444 00:22:14,770 --> 00:22:18,010 Speaker 2: future of AI really looks like. So, yes, you're right. Today, 445 00:22:18,030 --> 00:22:21,250 Speaker 2: we probably say we just don't need more than 100 446 00:22:21,250 --> 00:22:24,490 Speaker 2: IQ persons sitting in our pocket, managing our emails, sending 447 00:22:24,690 --> 00:22:27,970 Speaker 2: angry messages to the car rental company because they charged 448 00:22:27,990 --> 00:22:31,180 Speaker 2: me $ 2, 000 for a ding. I do think the consumer 449 00:22:31,220 --> 00:22:34,679 Speaker 2: layer AI is going to be, people ask me all 450 00:22:34,700 --> 00:22:36,780 Speaker 2: the time, like, what will get you really constructive on 451 00:22:36,820 --> 00:22:39,340 Speaker 2: AI again? I said, when we figure out that killer 452 00:22:39,380 --> 00:22:41,800 Speaker 2: consumer application layer. Because at the end of the day, 453 00:22:41,820 --> 00:22:45,159 Speaker 2: the consumer is everything. It's the economy. It's where, to 454 00:22:45,180 --> 00:22:47,840 Speaker 2: the point about revenues, like enterprise revenues, I just don't 455 00:22:47,900 --> 00:22:50,340 Speaker 2: think is going to plug that gap. Really? I just 456 00:22:50,359 --> 00:22:51,000 Speaker 2: don't see how. 457 00:22:51,040 --> 00:22:53,280 Speaker 3: Yeah, no, you're right. The numbers. I don't see. 458 00:22:53,359 --> 00:22:55,480 Speaker 2: It can probably get there eventually, but it can't get 459 00:22:55,520 --> 00:22:57,530 Speaker 2: there in the time frame that the market wants to 460 00:22:57,609 --> 00:23:00,649 Speaker 2: see the ROI. And you need to bust into the 461 00:23:00,690 --> 00:23:05,830 Speaker 2: consumer layer. And so back to the point about frontier, 462 00:23:05,850 --> 00:23:09,070 Speaker 2: the consumer probably doesn't need frontier for the time being 463 00:23:09,550 --> 00:23:14,290 Speaker 2: to do all of these things. And so my big 464 00:23:14,330 --> 00:23:17,369 Speaker 2: mantra is that one of the best technological changes that 465 00:23:17,410 --> 00:23:21,770 Speaker 2: we experience as humans is actually not necessarily some new 466 00:23:21,830 --> 00:23:26,639 Speaker 2: advanced tech. It's changing the application of existing tech. And 467 00:23:26,760 --> 00:23:29,520 Speaker 2: so it's very possible that we take the 100 IQ 468 00:23:29,680 --> 00:23:33,900 Speaker 2: agentic workflow that still kind of becomes almost ubiquitous within 469 00:23:33,920 --> 00:23:35,680 Speaker 2: the next 12 to 18 months. And we find really 470 00:23:35,720 --> 00:23:38,480 Speaker 2: cool ways to use it. And I think that's probably 471 00:23:38,960 --> 00:23:41,580 Speaker 2: where you're really going to get the next upswing in AI. 472 00:23:42,020 --> 00:23:45,560 Speaker 3: Can I just say, by the way, Tracy, last several days, 473 00:23:46,100 --> 00:23:48,740 Speaker 3: my wife was traveling out of town. My kids are 474 00:23:48,859 --> 00:23:52,399 Speaker 3: at a very easy age. It's not a big challenge 475 00:23:52,460 --> 00:23:54,930 Speaker 3: or anything to solo parent for a few days. But 476 00:23:55,280 --> 00:23:57,100 Speaker 3: it happened to be a very busy schedule for them, 477 00:23:57,170 --> 00:24:00,500 Speaker 3: birthday parties and a few other things. And it was 478 00:24:00,600 --> 00:24:05,120 Speaker 3: very nice having ChatGPT being able to say, where should 479 00:24:05,140 --> 00:24:06,540 Speaker 3: I be in the next hour? Where should I be 480 00:24:06,560 --> 00:24:08,600 Speaker 3: in the hour? Because there were a lot of things 481 00:24:08,640 --> 00:24:12,560 Speaker 3: going on. And it was very easy. It was very nice. 482 00:24:12,780 --> 00:24:14,320 Speaker 1: I saw your tweet on this, Joe. 483 00:24:14,359 --> 00:24:15,439 Speaker 3: No, you saw the tweet on the right. 484 00:24:15,460 --> 00:24:17,359 Speaker 1: This is part of the problem. You tweet everything before 485 00:24:17,400 --> 00:24:19,139 Speaker 1: you tell it to me, and so I can't react 486 00:24:19,160 --> 00:24:20,100 Speaker 1: to it in real time. 487 00:24:20,500 --> 00:24:22,640 Speaker 3: And now I'm really upset that you said that because 488 00:24:22,680 --> 00:24:24,920 Speaker 3: you saw the tweet from my private alt. I did 489 00:24:24,960 --> 00:24:27,350 Speaker 3: not tweet that from my main account. And now a 490 00:24:27,369 --> 00:24:28,649 Speaker 3: bunch of people are going to know that I have 491 00:24:28,690 --> 00:24:30,010 Speaker 3: an alt and they're going to ask me for it 492 00:24:30,030 --> 00:24:31,510 Speaker 3: and you're not going to get it. So I wish 493 00:24:31,530 --> 00:24:32,770 Speaker 3: you hadn't said that. Sorry. 494 00:24:33,070 --> 00:24:35,450 Speaker 1: Sorry. I can't, I can't differentiate the two. 495 00:24:35,609 --> 00:24:36,679 Speaker 2: Can I, can I, can I jump in? 496 00:24:36,690 --> 00:24:37,430 Speaker 1: Yeah, of course. 497 00:24:37,470 --> 00:24:40,790 Speaker 2: So what, what's that worth to you a month? What? Well, 498 00:24:40,869 --> 00:24:44,320 Speaker 2: just to have that ability, let's just say almost like 499 00:24:44,380 --> 00:24:48,080 Speaker 2: an account executive or something or like an administrator, you know, 500 00:24:48,630 --> 00:24:51,230 Speaker 2: Because to answer that question, it matters. It matters for 501 00:24:51,250 --> 00:24:52,530 Speaker 2: the revenue profile, right? 502 00:24:52,609 --> 00:24:54,379 Speaker 1: I mean, also, if you just had the right user 503 00:24:54,420 --> 00:24:58,290 Speaker 1: interface on a calendar, this would be solved too, wouldn't it? 504 00:24:59,060 --> 00:25:00,780 Speaker 1: If you know you have to be somewhere at 3 p.m., 505 00:25:00,780 --> 00:25:02,660 Speaker 1: 4 p.m., 5 p.m. There's a lot. 506 00:25:02,480 --> 00:25:05,350 Speaker 3: Of things that are in so many different emails. So, 507 00:25:05,369 --> 00:25:08,869 Speaker 3: for example, my daughter goes to– I take her to 508 00:25:08,910 --> 00:25:12,989 Speaker 3: play rehearsals. And some days, it's from 1 to 6 509 00:25:12,990 --> 00:25:15,690 Speaker 3: in the afternoon on Sundays. But some days, they don't 510 00:25:15,730 --> 00:25:17,770 Speaker 3: do the full cast for the five hours. So somewhere 511 00:25:17,810 --> 00:25:22,149 Speaker 3: in the email, it says, oh, these students or these participants, 512 00:25:22,350 --> 00:25:24,710 Speaker 3: they can leave at 3 and stuff. And it's scattered 513 00:25:24,750 --> 00:25:27,689 Speaker 3: across so many emails. And I don't know which emails 514 00:25:27,840 --> 00:25:31,500 Speaker 3: to read, which ones they're in. And so being able 515 00:25:31,520 --> 00:25:35,840 Speaker 3: to say, Chad, it's Sunday, October 25th or whatever. What 516 00:25:35,859 --> 00:25:37,540 Speaker 3: time am I supposed to pick up my daughter at 517 00:25:37,540 --> 00:25:40,359 Speaker 3: 3 or 6? It's very easy and nice to have. 518 00:25:40,920 --> 00:25:42,740 Speaker 1: Fair enough. I just want to go back to the 519 00:25:42,780 --> 00:25:45,800 Speaker 1: beginning of the conversation. You talked about neoclouds as the 520 00:25:45,840 --> 00:25:49,220 Speaker 1: sort of tip of the spear and maybe being in 521 00:25:49,300 --> 00:25:53,800 Speaker 1: the beginning of the denial phase. What do you think 522 00:25:53,859 --> 00:25:58,060 Speaker 1: would spook investors? What should we look out for next 523 00:25:58,220 --> 00:26:02,360 Speaker 1: in terms of a degradation in the AI bull belief? 524 00:26:03,460 --> 00:26:04,460 Speaker 3: That's a great question. 525 00:26:06,100 --> 00:26:08,720 Speaker 2: Off the top of my head, I would say it 526 00:26:08,760 --> 00:26:13,780 Speaker 2: would be some level of capex being reined in by 527 00:26:14,040 --> 00:26:16,460 Speaker 2: one of the hyperscalers. In fact, on the walk over here, 528 00:26:16,530 --> 00:26:20,409 Speaker 2: I was asked by one of my clients who was 529 00:26:21,390 --> 00:26:23,450 Speaker 2: questioning whether that would actually even come up potentially in 530 00:26:23,830 --> 00:26:27,970 Speaker 2: the results this week. And famous last words, it's probably 531 00:26:28,010 --> 00:26:31,840 Speaker 2: still premature. Joe, to answer your question about West Coast, 532 00:26:31,920 --> 00:26:35,090 Speaker 2: East Coast, if you think about West Coast, the further 533 00:26:35,130 --> 00:26:37,890 Speaker 2: you are closer to the epicenter of this technology, I'm sorry, 534 00:26:37,910 --> 00:26:40,530 Speaker 2: the closer you are to the epicenter, the more I 535 00:26:40,550 --> 00:26:44,970 Speaker 2: would say that both the companies and investors are bought in. 536 00:26:44,970 --> 00:26:45,820 Speaker 3: On that narrative. 537 00:26:46,500 --> 00:26:49,060 Speaker 2: And there's definitely kind of almost like an echo chamber 538 00:26:49,480 --> 00:26:52,780 Speaker 2: within that certain ecosystem. You think about the rooms that 539 00:26:52,800 --> 00:26:58,440 Speaker 2: people are sitting in. So my suspicion is that it's 540 00:26:58,520 --> 00:27:01,110 Speaker 2: not going to be these set of results. And that 541 00:27:01,170 --> 00:27:02,870 Speaker 2: may be we have to wait till the new year 542 00:27:03,400 --> 00:27:08,880 Speaker 2: that we get some kind of potential temperament of the enthusiasm. 543 00:27:10,260 --> 00:27:13,560 Speaker 2: But Google kind of set this all off again for 544 00:27:13,580 --> 00:27:16,040 Speaker 2: the next wave because on the CapEx side, it was 545 00:27:16,100 --> 00:27:19,560 Speaker 2: actually the last set of results where they said it's 546 00:27:19,580 --> 00:27:21,620 Speaker 2: going to be, I forget the exact wording they used, 547 00:27:21,660 --> 00:27:26,699 Speaker 2: but they were talking about CapEx being significantly higher. And 548 00:27:26,740 --> 00:27:29,000 Speaker 2: that got everyone enthused once again and got them all 549 00:27:29,040 --> 00:27:29,480 Speaker 2: G'd up. 550 00:27:29,970 --> 00:27:33,170 Speaker 1: Well, Oracle just said at its investor day that it 551 00:27:33,210 --> 00:27:35,650 Speaker 1: was going to limit its additional fundraising to, I think, 552 00:27:35,650 --> 00:27:41,240 Speaker 1: $ 20 billion through May. So we might be getting close 553 00:27:41,280 --> 00:27:43,720 Speaker 1: to that. But again, the reaction so far is this 554 00:27:43,780 --> 00:27:46,460 Speaker 1: is a good thing and the build-out can continue as 555 00:27:46,500 --> 00:27:49,140 Speaker 1: long as people are raising a reasonable amount of capital, 556 00:27:49,200 --> 00:27:49,520 Speaker 1: I guess. 557 00:27:49,920 --> 00:27:51,720 Speaker 2: Yeah. And that kind of goes back to maybe just 558 00:27:51,740 --> 00:27:55,369 Speaker 2: a slightly more holistic market view, which is This almost 559 00:27:55,850 --> 00:27:59,300 Speaker 2: S & P does nothing because what's good for hyperscalers 560 00:27:59,340 --> 00:28:02,139 Speaker 2: will be bad for memory. So you just get this 561 00:28:02,260 --> 00:28:06,040 Speaker 2: constant violent rotation. And I call it like a knife 562 00:28:06,080 --> 00:28:08,120 Speaker 2: fight in a phone booth, you know, for all these 563 00:28:08,220 --> 00:28:10,140 Speaker 2: funds that have to run long, short portfolios. 564 00:28:10,220 --> 00:28:12,400 Speaker 3: It's a nightmare. I'm glad you asked this because this 565 00:28:12,440 --> 00:28:14,389 Speaker 3: is a question that I wanted to ask you. So 566 00:28:14,410 --> 00:28:16,110 Speaker 3: it's like we look at the core weaves and there's 567 00:28:16,130 --> 00:28:19,889 Speaker 3: this anxiety and we see all the volatility. We're seeing 568 00:28:19,930 --> 00:28:24,209 Speaker 3: volatility everywhere except our measures of volatility. So if we 569 00:28:24,230 --> 00:28:27,550 Speaker 3: look at realized volatility, it's pretty low. If we look 570 00:28:27,570 --> 00:28:30,810 Speaker 3: at the VIX, it's not that high. Can you explain 571 00:28:31,030 --> 00:28:34,379 Speaker 3: why it is that there's all this anxiety about certain 572 00:28:34,410 --> 00:28:37,360 Speaker 3: parts of the most important trade in the entire world? 573 00:28:37,760 --> 00:28:39,800 Speaker 3: And yet, when we actually look at the way we 574 00:28:39,840 --> 00:28:41,380 Speaker 3: measure volatility, it looks very benign. 575 00:28:41,540 --> 00:28:44,500 Speaker 2: Yeah, I've been asked, this is probably the most asked 576 00:28:44,520 --> 00:28:47,260 Speaker 2: question that I get in some way, shape, or form. 577 00:28:48,000 --> 00:28:50,800 Speaker 2: And there's two parts to the answer, right? There's a 578 00:28:50,860 --> 00:28:53,770 Speaker 2: structural part. And then there's a really technical part. 579 00:28:53,790 --> 00:28:54,410 Speaker 3: Yeah, let's hear it. 580 00:28:54,590 --> 00:28:54,810 Speaker 1: All right. 581 00:28:54,830 --> 00:28:57,540 Speaker 2: So the structural part has got to do with the 582 00:28:57,620 --> 00:29:00,840 Speaker 2: fact with the way that money is deployed these days. 583 00:29:00,880 --> 00:29:04,560 Speaker 2: So obviously, equities are meant to be a long-only product, really. 584 00:29:04,940 --> 00:29:09,220 Speaker 2: But in this day and age, they're not. We've got 2020, 585 00:29:09,220 --> 00:29:13,720 Speaker 2: the AUM deployed in multi-manager funds was about $ 300 billion. 586 00:29:13,920 --> 00:29:18,670 Speaker 2: And today, it's in excess of a trillion dollars. And 587 00:29:18,770 --> 00:29:20,430 Speaker 2: if you think about the leverage- This is the pod shop, seriously. 588 00:29:21,290 --> 00:29:21,530 Speaker 1: Yeah. 589 00:29:21,610 --> 00:29:24,030 Speaker 2: I mean, not, not just pods, but obviously let's just 590 00:29:24,050 --> 00:29:25,830 Speaker 2: think about like that sort of ilk. 591 00:29:26,110 --> 00:29:26,270 Speaker 1: Yeah. 592 00:29:26,950 --> 00:29:29,490 Speaker 2: And it makes perfect sense, right? Because if you, especially 593 00:29:29,530 --> 00:29:31,410 Speaker 2: in a higher interest rate environment, high interest rate environment 594 00:29:31,430 --> 00:29:33,870 Speaker 2: is perfect if you're running a long short strategy, because 595 00:29:34,130 --> 00:29:36,090 Speaker 2: I deploy money on my gross long side and on 596 00:29:36,130 --> 00:29:38,500 Speaker 2: my growth short side. And actually I'm not really deploying. 597 00:29:38,520 --> 00:29:42,460 Speaker 2: I'm deploying units of risk, but the actual capital being 598 00:29:42,480 --> 00:29:45,620 Speaker 2: deployed is really only what I'm actually posting as margin 599 00:29:45,760 --> 00:29:47,700 Speaker 2: on my long and my short member. I can leverage 600 00:29:47,740 --> 00:29:51,720 Speaker 2: that too, through the prime prime businesses. So actually what 601 00:29:51,740 --> 00:29:53,680 Speaker 2: I'm doing is I'm taking a large chunk of capital 602 00:29:54,180 --> 00:29:56,730 Speaker 2: and I can put that into say a high interest bearing, 603 00:29:56,750 --> 00:30:00,550 Speaker 2: a high interest instrument. Mm-hmm clip a very nice coupon. 604 00:30:00,570 --> 00:30:03,030 Speaker 2: And then I try and generate alpha on that gross 605 00:30:03,070 --> 00:30:05,690 Speaker 2: capital deployed. That's the business model. And it's for the 606 00:30:05,730 --> 00:30:09,090 Speaker 2: most part, a very good business model, right? The problem is, 607 00:30:09,630 --> 00:30:13,090 Speaker 2: is that you've effectively tripled more than tripled the AUM 608 00:30:13,390 --> 00:30:17,219 Speaker 2: plus obviously leverage. And the number of securities you can 609 00:30:17,260 --> 00:30:20,220 Speaker 2: trade has stayed roughly the same or actually shrunk over time. 610 00:30:21,320 --> 00:30:24,610 Speaker 2: So that creates crowding out. So how do I how 611 00:30:24,630 --> 00:30:26,610 Speaker 2: do I organize or how do I mitigate that crowding 612 00:30:26,630 --> 00:30:29,540 Speaker 2: out effect? Well, I've got all kinds of cool risk parameters. So, 613 00:30:30,040 --> 00:30:33,420 Speaker 2: you know, Barra, I'm sure you've heard, or Axioma, just, 614 00:30:33,440 --> 00:30:37,959 Speaker 2: you know, all these different style factors, quality, value, size, growth, 615 00:30:37,980 --> 00:30:39,980 Speaker 2: you know, all these different things. You would have heard 616 00:30:40,040 --> 00:30:42,540 Speaker 2: us talk about them quite a lot. They've become ubiquitous 617 00:30:42,580 --> 00:30:44,820 Speaker 2: and obviously funds spend an inordinate amount of time and 618 00:30:44,860 --> 00:30:49,860 Speaker 2: computing power trying to neutralize those factors. So that's cool. 619 00:30:49,940 --> 00:30:52,420 Speaker 2: And you try and isolate the idiosyncratic risk of whatever 620 00:30:52,440 --> 00:30:56,190 Speaker 2: you're trying to get. Now, The problem is, is that 621 00:30:56,970 --> 00:30:58,630 Speaker 2: at the end of the day, everyone's still doing this, 622 00:30:59,030 --> 00:31:01,970 Speaker 2: which means all you're creating is different kinds of crowding. 623 00:31:03,630 --> 00:31:06,810 Speaker 2: And so that's why the volatility isn't showing up so 624 00:31:06,850 --> 00:31:10,570 Speaker 2: much at an index level, right? Because ultimately a lot 625 00:31:10,610 --> 00:31:14,560 Speaker 2: of this AUM is kind of index agnostic, so to speak. 626 00:31:15,540 --> 00:31:20,060 Speaker 2: And you're seeing the violent factor rotations happening because ultimately 627 00:31:20,080 --> 00:31:23,520 Speaker 2: you've got a huge amount of crowding sitting there. The 628 00:31:23,640 --> 00:31:26,100 Speaker 2: one thing that leads me into sort of the technical aspect, 629 00:31:26,140 --> 00:31:27,920 Speaker 2: which is the one thing that I think a lot 630 00:31:27,960 --> 00:31:30,900 Speaker 2: of funds have not yet been able to solve is 631 00:31:31,040 --> 00:31:33,160 Speaker 2: the AI factor, right? 632 00:31:33,960 --> 00:31:34,640 Speaker 3: Because you could be. 633 00:31:34,660 --> 00:31:36,520 Speaker 1: Everything either is or isn't AI now. 634 00:31:36,560 --> 00:31:37,680 Speaker 3: Right, exactly. 635 00:31:37,810 --> 00:31:40,430 Speaker 2: And this is a problem, right? It's dawned on me 636 00:31:40,470 --> 00:31:43,250 Speaker 2: recently because I've been asked a series of questions by 637 00:31:43,310 --> 00:31:47,770 Speaker 2: investors specifically along the lines of, have you managed to 638 00:31:47,830 --> 00:31:50,610 Speaker 2: map the AI factor? And the answer is no. Because 639 00:31:50,660 --> 00:31:52,950 Speaker 2: the AI factor is changing all of the time. The 640 00:31:52,990 --> 00:31:54,990 Speaker 2: betas are changing all of the time. Remember the beginning 641 00:31:55,010 --> 00:31:56,469 Speaker 2: of the year, you had a lot of people that 642 00:31:56,490 --> 00:32:00,230 Speaker 2: wanted to be short software. Now software is trading close 643 00:32:00,250 --> 00:32:03,170 Speaker 2: to 12 month highs. And so the idea of being 644 00:32:03,230 --> 00:32:06,370 Speaker 2: short software, it was part of the AI factor. And 645 00:32:06,430 --> 00:32:09,780 Speaker 2: now is it so much? Well, not necessarily unclear. So 646 00:32:09,840 --> 00:32:12,520 Speaker 2: because the narrative is shifting, the factor of AI is 647 00:32:12,580 --> 00:32:16,020 Speaker 2: shifting too. And that in turn complicates the ability to 648 00:32:16,140 --> 00:32:18,220 Speaker 2: funds to actually try and hedge that. And the bottom 649 00:32:18,240 --> 00:32:22,550 Speaker 2: line is, is you're either long AI or you're short AI, right? 650 00:32:22,630 --> 00:32:24,870 Speaker 2: And let's face it, the past three years, being short 651 00:32:24,930 --> 00:32:26,950 Speaker 2: AI probably hasn't been a particularly lucrative trade. 652 00:32:27,390 --> 00:32:30,690 Speaker 3: By the way, Tracy, in the time that Alex was 653 00:32:30,730 --> 00:32:33,430 Speaker 3: giving us, that Alty was giving us this great answer. 654 00:32:33,450 --> 00:32:36,310 Speaker 1: You sorted out your entire weekend schedule using ChatGPT. 655 00:32:36,410 --> 00:32:39,250 Speaker 3: No, in that time, we just got earnings from Alphabet 656 00:32:39,270 --> 00:32:42,890 Speaker 3: and Microsoft. And it's very interesting. Alphabet down 7.9%, revenue 657 00:32:42,910 --> 00:32:50,010 Speaker 3: X-TAC misses estimates. Microsoft jumped 6.4% after Q1 cloud revenue 658 00:32:50,050 --> 00:32:53,350 Speaker 3: tops estimate. So you get two gigantic stocks. They both 659 00:32:53,390 --> 00:32:56,890 Speaker 3: have very violent reactions. But if you're the indexed investor 660 00:32:56,910 --> 00:32:59,170 Speaker 3: on net, looks like it all kind of washes out. 661 00:32:59,340 --> 00:33:00,220 Speaker 1: Yeah, it does indeed. 662 00:33:00,360 --> 00:33:02,480 Speaker 3: But man, some of these, like these numbers are incredible. 663 00:33:02,860 --> 00:33:09,850 Speaker 3: Microsoft Q1 Azure Revenue X, Azure. This is their cloud platform. Azure. 664 00:33:10,070 --> 00:33:13,180 Speaker 3: What's your guess? You're not looking at a terminal the 665 00:33:13,220 --> 00:33:15,240 Speaker 3: way I am. What do you think it got? This 666 00:33:15,300 --> 00:33:15,620 Speaker 3: is fun. 667 00:33:15,700 --> 00:33:19,340 Speaker 2: The bokeh was, I think, 47% growth. 668 00:33:20,200 --> 00:33:22,600 Speaker 3: Yeah, man, you're really good at your job. So the 669 00:33:22,620 --> 00:33:26,080 Speaker 3: official estimate that we had collated at Bloomberg was 44.8%. 670 00:33:26,080 --> 00:33:28,220 Speaker 3: They hit 48%. 48% growth for these gigantic companies. That's 671 00:33:28,280 --> 00:33:29,180 Speaker 3: a good number. 672 00:33:29,200 --> 00:33:29,920 Speaker 2: That's unbelievable. 673 00:33:35,350 --> 00:33:38,530 Speaker 3: Cloud revenue at Google was $ 30 billion. I'm not seeing 674 00:33:38,630 --> 00:33:40,790 Speaker 3: what the growth number was. 675 00:33:41,070 --> 00:33:42,030 Speaker 1: But these are unreal. 676 00:33:42,370 --> 00:33:44,810 Speaker 2: Yeah. I mean, they are. And this goes back to 677 00:33:44,850 --> 00:33:47,330 Speaker 2: the point about kind of AI winners and losers. If 678 00:33:47,370 --> 00:33:50,770 Speaker 2: you've got a fantastic legacy business and you're a cash 679 00:33:50,790 --> 00:33:56,040 Speaker 2: volcano from your legacy operating business, then if you want 680 00:33:56,080 --> 00:34:02,440 Speaker 2: to go and embark upon the largest CapEx experiment really 681 00:34:02,460 --> 00:34:06,080 Speaker 2: in the history of America, total non sequitur, by the way, 682 00:34:06,520 --> 00:34:09,160 Speaker 2: I was trying to work out because we're spending, if 683 00:34:09,200 --> 00:34:11,300 Speaker 2: you believe the CapEx numbers and the terms of the forward, 684 00:34:11,340 --> 00:34:13,540 Speaker 2: and I'm sure that they're going to change from what 685 00:34:13,560 --> 00:34:16,069 Speaker 2: the consensus looks like now, but on face value, it's 686 00:34:16,110 --> 00:34:16,609 Speaker 2: about 12% of. 687 00:34:16,830 --> 00:34:17,870 Speaker 3: GDP, right? 688 00:34:18,350 --> 00:34:24,390 Speaker 2: It's like two Netherlands, right? Um, and I was thinking 689 00:34:25,430 --> 00:34:27,970 Speaker 2: you have not in modern history, like the Louisiana purchase, 690 00:34:27,989 --> 00:34:30,030 Speaker 2: the railroads, it's blown all that out of the water. 691 00:34:30,469 --> 00:34:32,739 Speaker 2: What have we done as a human society where we've 692 00:34:32,820 --> 00:34:35,529 Speaker 2: spent that much money? Certainly in a short period of time. 693 00:34:35,930 --> 00:34:38,230 Speaker 2: And I had to start getting really creative. I'm sure, Joe, 694 00:34:38,250 --> 00:34:40,989 Speaker 2: you'll be awake at night thinking about this and trying 695 00:34:41,010 --> 00:34:44,870 Speaker 2: to find some historical work on it. Pyramids, maybe? But 696 00:34:44,890 --> 00:34:48,009 Speaker 2: they took hundreds of years, if not thousands. Great Wall 697 00:34:48,050 --> 00:34:51,790 Speaker 2: of China, maybe? I can't think of a period where 698 00:34:51,850 --> 00:34:52,250 Speaker 2: we've spent. 699 00:34:52,270 --> 00:34:54,650 Speaker 3: All right, wait here. What do you think? How fast 700 00:34:54,670 --> 00:34:57,729 Speaker 3: do you think Google Cloud revenue grew? Again, I'm turning 701 00:34:57,750 --> 00:35:00,650 Speaker 3: my laptop away from me. I actually don't even... I 702 00:35:00,650 --> 00:35:03,070 Speaker 3: can't even believe this number is real. I'm giving you 703 00:35:03,090 --> 00:35:07,690 Speaker 3: a hint. 30-something percent? I don't really understand. I don't, I, 704 00:35:07,750 --> 00:35:08,190 Speaker 3: I almost. 705 00:35:09,110 --> 00:35:09,270 Speaker 1: Yeah. 706 00:35:09,270 --> 00:35:14,150 Speaker 3: 103% Google cloud revenue grew 103%. It seems great. I 707 00:35:14,170 --> 00:35:16,910 Speaker 3: don't know. All these numbers are insane. Anyway, that's what 708 00:35:16,930 --> 00:35:17,280 Speaker 3: it says. 709 00:35:18,360 --> 00:35:21,879 Speaker 2: I mean, these are impressive businesses and this is, this 710 00:35:21,940 --> 00:35:23,320 Speaker 2: is why, you know, a lot of people say S 711 00:35:23,360 --> 00:35:24,200 Speaker 2: and P undefeated. 712 00:35:24,260 --> 00:35:24,440 Speaker 1: Right. 713 00:35:24,460 --> 00:35:26,600 Speaker 2: I mean, by the way, funny, the concentration of the 714 00:35:26,680 --> 00:35:29,469 Speaker 2: S and P it basically given Tracy, to your point, 715 00:35:29,530 --> 00:35:33,580 Speaker 2: everything's kind of AI. Concentration of S & P, next year, 75% 716 00:35:33,580 --> 00:35:34,859 Speaker 2: of your earnings growth in the S & P is 717 00:35:34,890 --> 00:35:38,850 Speaker 2: going to come from memory, from semiconductors. 718 00:35:38,950 --> 00:35:39,730 Speaker 3: What percent? 719 00:35:40,430 --> 00:35:46,090 Speaker 2: Sorry, a bigger part. It's actually, it's about two-thirds from semiconductors, 720 00:35:46,170 --> 00:35:48,780 Speaker 2: and then the broader AI ecosystem gets you to around 721 00:35:48,800 --> 00:35:51,660 Speaker 2: about 75%. Wow. So you basically, S & P has 722 00:35:51,680 --> 00:35:52,260 Speaker 2: become Cosby. 723 00:35:53,950 --> 00:35:56,570 Speaker 3: Oh, cospification. 724 00:35:57,010 --> 00:35:59,910 Speaker 1: Cospification, yeah. All right, well, Alex, we're going to leave 725 00:35:59,950 --> 00:36:01,569 Speaker 1: it there because we want to get this episode out 726 00:36:01,610 --> 00:36:03,670 Speaker 1: pretty quickly. But thank you so much for coming back 727 00:36:03,730 --> 00:36:04,210 Speaker 1: on All Thoughts. 728 00:36:04,250 --> 00:36:05,330 Speaker 3: Really appreciate it. 729 00:36:05,350 --> 00:36:05,690 Speaker 2: Thank you. 730 00:36:05,790 --> 00:36:07,550 Speaker 3: See you next week. Thank you so much. 731 00:36:21,540 --> 00:36:23,319 Speaker 1: Joe, it was great catching up with Alex. I think 732 00:36:23,360 --> 00:36:25,540 Speaker 1: he does a great job of sort of synthesizing a 733 00:36:25,580 --> 00:36:27,780 Speaker 1: lot of different things going on in the market right now. 734 00:36:28,219 --> 00:36:31,279 Speaker 1: But I do think the key thing that's emerging is 735 00:36:31,340 --> 00:36:35,189 Speaker 1: this differentiation idea and the idea that we're moving from 736 00:36:35,210 --> 00:36:39,450 Speaker 1: this sort of capital abundant environment into an area where 737 00:36:39,469 --> 00:36:41,870 Speaker 1: people are perhaps a little bit more selective and what 738 00:36:41,910 --> 00:36:43,319 Speaker 1: that means for the business models. 739 00:36:44,020 --> 00:36:49,339 Speaker 3: I do think, completely agree. I also think, you know, every... 740 00:36:49,770 --> 00:36:53,870 Speaker 3: The classic is different this time is, oh, interest rates 741 00:36:53,930 --> 00:36:56,910 Speaker 3: don't matter anymore, right? And this is a classic thing. Oh, 742 00:36:56,950 --> 00:36:59,230 Speaker 3: this is not going to be a rate-sensitive industry because 743 00:36:59,290 --> 00:37:03,969 Speaker 3: the numbers are so unbelievably high. If you're growing 44% 744 00:37:03,969 --> 00:37:06,089 Speaker 3: in a year, what do you care about when the 745 00:37:06,180 --> 00:37:09,719 Speaker 3: rates are 5% or 5.25? It's completely irrelevant. You're just 746 00:37:09,739 --> 00:37:12,120 Speaker 3: going to keep building, et cetera. But that is a 747 00:37:12,180 --> 00:37:16,060 Speaker 3: lie that investors tell themselves. Rates, at least at the margin, 748 00:37:16,460 --> 00:37:19,600 Speaker 3: at least at the tip of the sphere– sphere– at 749 00:37:19,660 --> 00:37:23,620 Speaker 3: least for some of these leveraged companies like the Neoclouds 750 00:37:23,640 --> 00:37:26,759 Speaker 3: or whatever, for one thing, it looks like a rates matter. 751 00:37:26,960 --> 00:37:30,950 Speaker 1: Yeah. And also, I mean, just the costification idea of 752 00:37:31,030 --> 00:37:33,489 Speaker 1: the market, it is really quite stunning because a lot 753 00:37:33,510 --> 00:37:37,330 Speaker 1: of the earnings power of the tech companies, the hyperscalers, 754 00:37:37,410 --> 00:37:39,529 Speaker 1: is predicated on the fact that, well, other companies are 755 00:37:39,550 --> 00:37:41,629 Speaker 1: going to buy the services so that they can boost 756 00:37:41,670 --> 00:37:43,650 Speaker 1: their own productivity and get the earnings up. 757 00:37:44,070 --> 00:37:44,390 Speaker 2: But like. 758 00:37:45,180 --> 00:37:48,440 Speaker 1: we're not actually seeing that materialize in their earnings numbers 759 00:37:48,700 --> 00:37:51,580 Speaker 1: just yet. And in the meantime, everything, the whole index 760 00:37:51,620 --> 00:37:54,700 Speaker 1: just becomes AI more and more on the expectation that 761 00:37:54,739 --> 00:37:57,240 Speaker 1: eventually it won't just be about AI. 762 00:37:57,739 --> 00:38:00,239 Speaker 3: This is the thing that like, you know, when you 763 00:38:00,280 --> 00:38:03,680 Speaker 3: talked about the pod shops, for example, the whole premise 764 00:38:03,719 --> 00:38:06,200 Speaker 3: of a pod shop is that like everyone is supposed 765 00:38:06,219 --> 00:38:08,540 Speaker 3: to find their own undifferentiated alpha. 766 00:38:08,620 --> 00:38:11,720 Speaker 1: Yeah. Where is that right now? 767 00:38:11,760 --> 00:38:15,880 Speaker 3: So I think like, basically, if you're an investor, And 768 00:38:15,890 --> 00:38:18,460 Speaker 3: you made money in the last year at scale. I mean, 769 00:38:18,480 --> 00:38:20,279 Speaker 3: of course, like you might be able to like, oh, 770 00:38:20,340 --> 00:38:23,360 Speaker 3: I opened, you know, a new like pasta restaurant in 771 00:38:23,380 --> 00:38:25,480 Speaker 3: the East Village. That's a separate question. But if you're 772 00:38:25,580 --> 00:38:28,879 Speaker 3: at scale and you made money in the last year, 773 00:38:29,260 --> 00:38:31,620 Speaker 3: in some way it was connected to AI. You have 774 00:38:31,640 --> 00:38:34,560 Speaker 3: all these different pods and some of them are chips 775 00:38:34,620 --> 00:38:37,299 Speaker 3: and some of them are industrials, et cetera. But it's 776 00:38:37,380 --> 00:38:39,430 Speaker 3: hard for me to believe that the ones who are 777 00:38:39,469 --> 00:38:43,390 Speaker 3: making money aren't just in some way finding a way 778 00:38:43,430 --> 00:38:44,529 Speaker 3: to belong to AI fans. 779 00:38:45,000 --> 00:38:45,700 Speaker 1: Levered long. 780 00:38:45,719 --> 00:38:46,180 Speaker 3: Yeah, levered long. 781 00:38:46,300 --> 00:38:47,580 Speaker 1: The AI factor, I would imagine. 782 00:38:47,719 --> 00:38:52,160 Speaker 3: The AI factor. You know, I was talking to... It's 783 00:38:52,180 --> 00:38:54,680 Speaker 3: so funny, you know, we try to... Alex has such 784 00:38:54,700 --> 00:38:56,700 Speaker 3: a good read on the industry, but also it's great 785 00:38:56,719 --> 00:38:58,469 Speaker 3: to talk to him because he talks to clients to 786 00:38:58,489 --> 00:39:00,210 Speaker 3: get the Wall Street view and so forth. I was 787 00:39:00,250 --> 00:39:04,350 Speaker 3: talking to this computer science student at Stanford recently, and 788 00:39:04,370 --> 00:39:06,109 Speaker 3: they were like, oh, Wall Street is an AI pill. 789 00:39:06,130 --> 00:39:07,569 Speaker 3: And I was like, what are you talking about, Wall 790 00:39:07,590 --> 00:39:09,650 Speaker 3: Street is an AI pill? Look at all these stocks 791 00:39:09,670 --> 00:39:10,730 Speaker 3: that are going to the moon. 792 00:39:10,810 --> 00:39:12,790 Speaker 1: The numbers on my screen tell a different story. 793 00:39:12,810 --> 00:39:13,569 Speaker 3: What are you talking about? 794 00:39:13,850 --> 00:39:14,110 Speaker 1: But. 795 00:39:15,110 --> 00:39:18,950 Speaker 3: That person's view was that if you're even having a 796 00:39:19,010 --> 00:39:24,480 Speaker 3: discussion about the possibility of overabundance of GPUs, you just 797 00:39:24,500 --> 00:39:27,160 Speaker 3: don't get it. like what do you well you're like 798 00:39:27,380 --> 00:39:30,219 Speaker 3: oh you you don't you don't get ai sufficiently it's 799 00:39:30,239 --> 00:39:31,880 Speaker 3: so funny it's. 800 00:39:32,040 --> 00:39:35,219 Speaker 1: Um yeah definitely an interesting time in markets shall we 801 00:39:35,239 --> 00:39:35,920 Speaker 1: leave it there. 802 00:39:36,160 --> 00:39:40,390 Speaker 3: Alphabet they do 32 billion tokens per minute that was 803 00:39:40,450 --> 00:39:42,610 Speaker 3: up 40 percent q of recovery i don't even know 804 00:39:42,630 --> 00:39:43,890 Speaker 3: what these words mean let's leave. 805 00:39:43,790 --> 00:39:46,400 Speaker 1: It there All right. This has been another episode of 806 00:39:46,440 --> 00:39:48,939 Speaker 1: the All Thoughts Podcast. I'm Traci Allaway. You can follow 807 00:39:48,960 --> 00:39:50,100 Speaker 1: me at Traci Allaway. 808 00:39:50,460 --> 00:39:53,430 Speaker 3: And I'm Jill Wiesenthal. You can follow me at The Stalwart. 809 00:39:53,570 --> 00:39:57,180 Speaker 3: Follow our producers, Carmen Rodriguez at Carmen Armand, Dashiell Bennett 810 00:39:57,239 --> 00:40:00,430 Speaker 3: at Dashbot, Kale Brooks at Kale Brooks, and Kevin Lozano 811 00:40:00,590 --> 00:40:03,170 Speaker 3: at Kevin Lloyd Lozano. You can chat about all these 812 00:40:03,210 --> 00:40:08,430 Speaker 3: topics 24-7 in our Discord, discord.gg slash oddlots. For more 813 00:40:08,510 --> 00:40:11,330 Speaker 3: oddlots content, check out our daily newsletter, which you can 814 00:40:11,370 --> 00:40:13,210 Speaker 3: find at glimmer.com slash oddlots. 815 00:40:13,940 --> 00:40:16,120 Speaker 1: And if you enjoyed this conversation, if you like it 816 00:40:16,280 --> 00:40:18,200 Speaker 1: when we check in on the markets, then please leave 817 00:40:18,239 --> 00:40:21,239 Speaker 1: us a positive review on your favorite podcast platform. And 818 00:40:21,260 --> 00:40:23,879 Speaker 1: if you're watching on video, then please like or leave 819 00:40:23,920 --> 00:40:25,880 Speaker 1: a comment or better yet, subscribe. 820 00:40:25,900 --> 00:40:29,020 Speaker 3: And remember, if you're a Bloomberg subscriber, you can listen 821 00:40:29,060 --> 00:40:32,660 Speaker 3: to all of our episodes absolutely ad-free. All you need 822 00:40:32,700 --> 00:40:35,270 Speaker 3: to do is find the Bloomberg channel on Apple Podcasts 823 00:40:35,370 --> 00:40:37,110 Speaker 3: and follow the instructions there. 824 00:40:37,150 --> 00:40:37,810 Speaker 1: Thanks for joining. 825 00:40:47,660 --> 00:40:48,450 Speaker 3: Thank you.