1 00:00:02,360 --> 00:00:09,640 Speaker 1: Bloomberg Audio Studios, Podcasts, radio News. I'm Stephen Carroll, and 2 00:00:09,680 --> 00:00:12,360 Speaker 1: this is Here's Why, where we take one news story 3 00:00:12,440 --> 00:00:14,560 Speaker 1: and explain it in just a few minutes with our 4 00:00:14,600 --> 00:00:16,240 Speaker 1: experts here at Bloomberg. 5 00:00:18,680 --> 00:00:21,639 Speaker 2: The tech Joan's disclose huge AI spending plans in their 6 00:00:21,720 --> 00:00:25,239 Speaker 2: latest earnings reports. Alphabet is kicking off a seven part 7 00:00:25,320 --> 00:00:28,280 Speaker 2: bond sale, adding to a borrowing spree by companies at 8 00:00:28,280 --> 00:00:30,520 Speaker 2: the forefront of the AI investment boom. 9 00:00:30,560 --> 00:00:32,800 Speaker 3: The idea of one hundred year bond from one of 10 00:00:32,800 --> 00:00:34,800 Speaker 3: these hyperskillers is crazy to me because we don't even 11 00:00:34,840 --> 00:00:36,680 Speaker 3: know what the next three years is going to look like, 12 00:00:36,800 --> 00:00:38,200 Speaker 3: let alone one hundred years. 13 00:00:38,120 --> 00:00:41,600 Speaker 1: You might be bracing for four hundred billion dollars of 14 00:00:41,720 --> 00:00:46,040 Speaker 1: investment grade debt coming to finance AI. The AI spending 15 00:00:46,080 --> 00:00:50,200 Speaker 1: spree is gaining speed, and it's coming with massive borrowing. 16 00:00:50,560 --> 00:00:54,080 Speaker 1: Bloomberg Intelligence estimates that the biggest tech firms will spend 17 00:00:54,120 --> 00:00:57,720 Speaker 1: more than four trillion dollars by twenty thirty in their 18 00:00:57,760 --> 00:01:01,840 Speaker 1: efforts to dominate the technology. To fund it, they're spending 19 00:01:01,920 --> 00:01:05,759 Speaker 1: cash and raising huge amounts of money in depth markets. 20 00:01:06,480 --> 00:01:13,440 Speaker 1: Here's why AI ambitions are fueling a borrowing boom. Our 21 00:01:13,480 --> 00:01:16,520 Speaker 1: market's live strategistic Skyler Montgomery Coning joins us. 22 00:01:16,560 --> 00:01:17,559 Speaker 2: Now for more. 23 00:01:17,959 --> 00:01:21,200 Speaker 1: Skyler, First of all, can you put this scale of 24 00:01:21,280 --> 00:01:22,880 Speaker 1: borrowing in context for us? 25 00:01:22,880 --> 00:01:24,680 Speaker 2: Have we seen anything like this before? 26 00:01:25,080 --> 00:01:27,680 Speaker 3: I mean, the short answer is no, we haven't seen 27 00:01:28,000 --> 00:01:31,880 Speaker 3: this kind of spending as well as boring before, at 28 00:01:31,959 --> 00:01:35,440 Speaker 3: least not in my lifetime. Now, the slightly longer answer 29 00:01:35,720 --> 00:01:41,120 Speaker 3: is that technological revelations or revolutions require investment, and investment 30 00:01:41,160 --> 00:01:44,279 Speaker 3: needs to be financed. But what makes this AI cycle 31 00:01:44,400 --> 00:01:48,560 Speaker 3: different is both the scale and the financing mix. Now, 32 00:01:48,560 --> 00:01:52,720 Speaker 3: the spending is exceptionally large, and it's increasingly debt finance. 33 00:01:53,120 --> 00:01:55,640 Speaker 3: From what I've read, more money has already been put 34 00:01:55,720 --> 00:01:59,320 Speaker 3: into this technology boom than telecom networks in the nineteen 35 00:01:59,400 --> 00:02:03,200 Speaker 3: nineties or anything that's come in recent decades, but less 36 00:02:03,200 --> 00:02:05,760 Speaker 3: than say, the railroads in the eighteen forties, which is 37 00:02:05,840 --> 00:02:08,400 Speaker 3: quite a while ago and was one of the largest 38 00:02:08,440 --> 00:02:12,720 Speaker 3: infrastructure booms in economic history. What makes AI unique is 39 00:02:12,720 --> 00:02:16,240 Speaker 3: that it combines a kind of breadth and capital intensity. 40 00:02:16,280 --> 00:02:19,160 Speaker 3: So the technology can be deployed across almost every sector 41 00:02:19,200 --> 00:02:22,560 Speaker 3: of the economy, but it's also capital intensive. It needs 42 00:02:22,639 --> 00:02:26,240 Speaker 3: long lived, power intensive assets to give you some numbers 43 00:02:26,280 --> 00:02:30,280 Speaker 3: on debts. Specifically, the Dallas Fed thinks that three hundred 44 00:02:30,280 --> 00:02:34,120 Speaker 3: billion of investment grade or IG issuance related to the 45 00:02:34,160 --> 00:02:37,560 Speaker 3: AI buildout is a reasonable estimate for this year. Now, 46 00:02:37,560 --> 00:02:40,520 Speaker 3: that's a very big shift for companies that have historically 47 00:02:40,560 --> 00:02:44,799 Speaker 3: been relatively capex light and cash flow funded. It's also 48 00:02:44,919 --> 00:02:48,880 Speaker 3: significantly for the credit market itself. That three hundred billion 49 00:02:48,960 --> 00:02:52,480 Speaker 3: value is roughly seventeen percent of total USIG issuance in 50 00:02:52,520 --> 00:02:54,480 Speaker 3: twenty twenty five, and that debt is. 51 00:02:54,440 --> 00:02:55,919 Speaker 2: Only being added for AI. 52 00:02:56,560 --> 00:02:59,959 Speaker 3: So we've seen technology booms before, but the combination of 53 00:03:00,200 --> 00:03:03,000 Speaker 3: scale and the high degree of debt financing makes this 54 00:03:03,120 --> 00:03:04,079 Speaker 3: cycle stand out. 55 00:03:04,480 --> 00:03:07,240 Speaker 1: Yeah, it's certainly remarkable in so many aspects. Can you 56 00:03:07,280 --> 00:03:09,440 Speaker 1: just remind us of what the tech companies are planning 57 00:03:09,480 --> 00:03:10,840 Speaker 1: on spending all this money on. 58 00:03:11,400 --> 00:03:14,560 Speaker 3: Yeah, some of its continued development, as you would expect. 59 00:03:14,600 --> 00:03:15,560 Speaker 2: So that's stuff like. 60 00:03:16,040 --> 00:03:20,680 Speaker 3: Improving large language models, training more advanced systems, building chips. 61 00:03:21,200 --> 00:03:24,400 Speaker 3: But the majority of the spending is far more physical 62 00:03:24,520 --> 00:03:27,760 Speaker 3: than I think people realize. It isn't just software, it's 63 00:03:27,880 --> 00:03:32,640 Speaker 3: kind of industrial scale infrastructure. Now companies are building enormous 64 00:03:32,760 --> 00:03:35,480 Speaker 3: data centers that are the size of kind of multiple 65 00:03:35,480 --> 00:03:39,520 Speaker 3: football fields. In other words, AI is becoming an energy 66 00:03:39,680 --> 00:03:41,200 Speaker 3: and infrastructure. 67 00:03:40,600 --> 00:03:43,280 Speaker 2: Story that has real world spillovers. 68 00:03:43,360 --> 00:03:47,120 Speaker 3: Now, data centers consume large amount of electricity and land, 69 00:03:47,200 --> 00:03:50,360 Speaker 3: which puts pressure on local grids and resources. It also 70 00:03:50,440 --> 00:03:54,560 Speaker 3: drives demand for construction and commodities. So this isn't a 71 00:03:54,640 --> 00:03:58,520 Speaker 3: typical tech investment cycle, and that's why the boring is 72 00:03:58,560 --> 00:04:01,960 Speaker 3: so large. These are long lived, capital intensive assets, not 73 00:04:02,080 --> 00:04:04,080 Speaker 3: kind of incremental software tweaks. 74 00:04:04,480 --> 00:04:07,320 Speaker 1: And is that also why we're seeing these companies which 75 00:04:07,320 --> 00:04:11,480 Speaker 1: are very rich, borrowing rather than using cash issuing shares. 76 00:04:12,160 --> 00:04:14,240 Speaker 3: Again, here there's kind of a short answer and a 77 00:04:14,280 --> 00:04:18,320 Speaker 3: long answer. The short answer is they can. There's clearly 78 00:04:18,360 --> 00:04:21,800 Speaker 3: strong demand for this debt. Investors want high quality yield, 79 00:04:22,360 --> 00:04:25,960 Speaker 3: and large tech issuers are seen as very safe borers. 80 00:04:26,880 --> 00:04:30,240 Speaker 3: The long answer is more about capital structure. So these 81 00:04:30,240 --> 00:04:33,320 Speaker 3: companies are coming from a position of strength. They have 82 00:04:33,480 --> 00:04:36,720 Speaker 3: robust balance sheets and strong cash flows, so it makes 83 00:04:36,760 --> 00:04:39,160 Speaker 3: sense to introduce more debt into the mix. 84 00:04:39,640 --> 00:04:40,359 Speaker 2: For corporates. 85 00:04:40,480 --> 00:04:43,840 Speaker 3: Debt financing is actually cheaper than equity finanting, So even 86 00:04:43,880 --> 00:04:47,200 Speaker 3: though we've had government bond yields rise in recent years, 87 00:04:47,680 --> 00:04:51,280 Speaker 3: high grade tech companies can borrow at relatively tight spreads, 88 00:04:51,760 --> 00:04:54,880 Speaker 3: so that lowers their weighted average cost of capital. 89 00:04:55,240 --> 00:04:57,279 Speaker 2: There's also a flexibility. 90 00:04:56,560 --> 00:04:59,359 Speaker 3: Element in that instruments like convertible bonds that some of 91 00:04:59,400 --> 00:05:02,760 Speaker 3: these companies issuing gives them the option to convert debt 92 00:05:02,800 --> 00:05:05,320 Speaker 3: to equity in the future. And this is also just 93 00:05:05,320 --> 00:05:08,240 Speaker 3: a very large investment cycle, so funding it with debt 94 00:05:08,360 --> 00:05:12,480 Speaker 3: rather than equities means you aren't diluting shareholders or running 95 00:05:12,480 --> 00:05:14,159 Speaker 3: down cash to the point that you need to worry 96 00:05:14,200 --> 00:05:19,280 Speaker 3: about liquidity. So ultimately it's about capital optimization. One debt 97 00:05:19,320 --> 00:05:21,560 Speaker 3: is cheaper than equity in your balance sheet is healthy, 98 00:05:21,960 --> 00:05:24,560 Speaker 3: It can make sense to borrow to fund a long 99 00:05:24,640 --> 00:05:25,960 Speaker 3: term growth opportunity. 100 00:05:26,400 --> 00:05:29,560 Speaker 1: Okay, and even within the borrowing there's loads of firsts 101 00:05:29,560 --> 00:05:32,520 Speaker 1: here too, alphabout issuing Sterling and Swiss frank debt and 102 00:05:32,560 --> 00:05:35,040 Speaker 1: even a one hundred year note. What does it tell 103 00:05:35,120 --> 00:05:37,880 Speaker 1: us about the investor appetite for this kind of debt? 104 00:05:38,400 --> 00:05:40,400 Speaker 3: Well, I just want to say first that it's wild, 105 00:05:40,680 --> 00:05:43,160 Speaker 3: but also that it tells you that there's a lot 106 00:05:43,160 --> 00:05:46,120 Speaker 3: of demand for this debt on a global scale, and 107 00:05:46,160 --> 00:05:49,520 Speaker 3: that companies are trying to capture a wide investor base. 108 00:05:49,640 --> 00:05:53,400 Speaker 3: The one hundred year one in particular, is wild investors 109 00:05:53,440 --> 00:05:54,240 Speaker 3: are really. 110 00:05:54,120 --> 00:05:56,880 Speaker 1: At it where this technologist will be in ten minutes, 111 00:05:56,920 --> 00:05:58,520 Speaker 1: it feels like exactly. 112 00:05:58,600 --> 00:06:00,920 Speaker 3: And investors are essentially saying that they're willing to lock 113 00:06:01,000 --> 00:06:03,840 Speaker 3: up capital for what is likely to be longer than 114 00:06:03,880 --> 00:06:08,760 Speaker 3: their lifetimes at relatively modest spreads because they believe in 115 00:06:09,040 --> 00:06:12,720 Speaker 3: the durability in the growth potential of these companies. Now, 116 00:06:12,760 --> 00:06:15,400 Speaker 3: the fact that this issuance is happening in foreign currencies 117 00:06:15,440 --> 00:06:18,839 Speaker 3: you mentioned sterling in the Swiss franc that also highlights 118 00:06:18,880 --> 00:06:21,279 Speaker 3: the breadth of demand, So it's not just u as investors, 119 00:06:21,320 --> 00:06:24,800 Speaker 3: it's global. It also speaks to the scarcity in that 120 00:06:24,839 --> 00:06:27,680 Speaker 3: there isn't a huge supply of high grade credit with 121 00:06:27,760 --> 00:06:31,599 Speaker 3: structural growth exposure, and with the tech trade already crowded 122 00:06:31,640 --> 00:06:35,080 Speaker 3: in equity markets, some events ers are clearly looking for 123 00:06:35,400 --> 00:06:39,159 Speaker 3: alternative ways to gain exposure, and credit provides a different 124 00:06:39,279 --> 00:06:41,960 Speaker 3: entry point into the same kind of structural trade. 125 00:06:42,320 --> 00:06:45,080 Speaker 1: What does all of this delege of debt mean for 126 00:06:45,160 --> 00:06:46,720 Speaker 1: the credit markets more broadly? 127 00:06:47,400 --> 00:06:48,880 Speaker 2: This is the interesting part to me. 128 00:06:49,800 --> 00:06:52,800 Speaker 3: Whenever you get this structural shift within markets on such 129 00:06:52,800 --> 00:06:56,600 Speaker 3: a large scale that are bound to be unintended spillovers. 130 00:06:56,600 --> 00:07:00,320 Speaker 3: For credit markets, the surge and AI related issuance likely 131 00:07:00,320 --> 00:07:04,680 Speaker 3: has a couple. Mostly it's around diversification. So credit's often 132 00:07:04,720 --> 00:07:08,680 Speaker 3: held because it sits between equities and government bonds. It 133 00:07:08,760 --> 00:07:13,320 Speaker 3: offers income and some growth exposure, but with lower volatility 134 00:07:13,360 --> 00:07:19,000 Speaker 3: than stocks. However, as AI related issuance increases, that balance shifts. 135 00:07:19,360 --> 00:07:22,560 Speaker 3: So historically credit indices have had relatively low exposure to 136 00:07:22,640 --> 00:07:25,880 Speaker 3: tacks because these companies didn't need to borrow much. But 137 00:07:26,000 --> 00:07:29,920 Speaker 3: now with heavy AI cupax, large tech firms are issuing 138 00:07:29,960 --> 00:07:33,000 Speaker 3: substantial amounts of debt and that increases their weight in 139 00:07:33,120 --> 00:07:36,480 Speaker 3: credit induicy, so it makes the market more concentrated, and 140 00:07:36,520 --> 00:07:40,000 Speaker 3: it incrementally makes it more similar to equity benchmarks. Now 141 00:07:40,000 --> 00:07:43,440 Speaker 3: that's especially true if their issuance crowds out other sectors 142 00:07:43,480 --> 00:07:47,120 Speaker 3: from selling bonds. At the same time, spreads are already 143 00:07:47,240 --> 00:07:51,200 Speaker 3: very tight, so the risk bruffer that credit provides over 144 00:07:51,240 --> 00:07:54,840 Speaker 3: government bonds is already limited, and that means returns become 145 00:07:54,920 --> 00:07:58,840 Speaker 3: more dependent on moves and underlying rates rather than spread compression. 146 00:07:59,360 --> 00:08:01,640 Speaker 3: And that's particular really important because much of this tech 147 00:08:01,640 --> 00:08:05,000 Speaker 3: issueance is long duration. What I mean by duration, it's 148 00:08:05,080 --> 00:08:07,800 Speaker 3: really just a measure of how sensitive a bond's price 149 00:08:07,880 --> 00:08:11,720 Speaker 3: is to interest rates and is related to maturity in 150 00:08:11,760 --> 00:08:14,560 Speaker 3: that you know, very simply, a longer maturity typically means 151 00:08:14,560 --> 00:08:17,960 Speaker 3: a higher duration, and the US Investment Grade Index already 152 00:08:17,960 --> 00:08:20,760 Speaker 3: has a duration of around seven years. But the issues 153 00:08:20,800 --> 00:08:22,760 Speaker 3: that we're seeing out of TECH is forty year one 154 00:08:22,840 --> 00:08:26,000 Speaker 3: hundred year bonds, and that will push the average duration 155 00:08:26,080 --> 00:08:28,720 Speaker 3: of the index higher. And that means you get an 156 00:08:28,720 --> 00:08:32,520 Speaker 3: increased sensitivity to interest rates and increased volatility of these 157 00:08:32,520 --> 00:08:36,560 Speaker 3: credit indices. So taken together, credit could become more equity 158 00:08:36,760 --> 00:08:40,839 Speaker 3: like in its concentration and more rate sensitive in its behavior, 159 00:08:41,320 --> 00:08:45,280 Speaker 3: which chips away the diversification benefits and visitors typically expect 160 00:08:45,320 --> 00:08:47,720 Speaker 3: from it now. As I said earlier, with any major 161 00:08:47,720 --> 00:08:51,280 Speaker 3: market shift, there are unintended consequences. These are the ones 162 00:08:51,320 --> 00:08:53,840 Speaker 3: we can see, but there are also others that will 163 00:08:53,920 --> 00:08:54,640 Speaker 3: likely emerge too. 164 00:08:55,080 --> 00:08:57,760 Speaker 1: Yeah. I mean, it's a fascinating rewiring of these markets 165 00:08:57,880 --> 00:09:00,680 Speaker 1: as well. Dare I ask the question what happens if 166 00:09:00,720 --> 00:09:02,800 Speaker 1: this all goes wrong? What if AI doesn't meet the 167 00:09:02,800 --> 00:09:04,319 Speaker 1: big expectations there are for it. 168 00:09:05,360 --> 00:09:06,760 Speaker 2: The main risk is simple. 169 00:09:06,800 --> 00:09:09,920 Speaker 3: It's just that the return of investment doesn't justify the 170 00:09:09,960 --> 00:09:12,640 Speaker 3: scale of spending. So in other words, it's just that 171 00:09:12,679 --> 00:09:16,840 Speaker 3: the investment is unprofitable. Now at the market level, concentration 172 00:09:17,240 --> 00:09:20,360 Speaker 3: amplifies that risk. So tech already represents a large share 173 00:09:20,360 --> 00:09:23,199 Speaker 3: of equity indices, and as I said, it's becoming a 174 00:09:23,320 --> 00:09:27,200 Speaker 3: larger share of credit indices as well. That means investors, 175 00:09:27,240 --> 00:09:31,480 Speaker 3: often passively are increasingly dependent on a relatively small group 176 00:09:31,559 --> 00:09:36,240 Speaker 3: of companies continuing to execute. If those companies disappoint, the 177 00:09:36,320 --> 00:09:40,400 Speaker 3: reprogressions aren't as contained because of their index waights. Weakness 178 00:09:40,400 --> 00:09:44,840 Speaker 3: in the sector weighs dispportionately on broader equity and credit markets, 179 00:09:45,320 --> 00:09:48,160 Speaker 3: and with credit markets in particular more involved. Now there's 180 00:09:48,160 --> 00:09:53,080 Speaker 3: also a financing angle where if returns disappoint, spreads widen, 181 00:09:53,559 --> 00:09:57,880 Speaker 3: and that means funding costs widen or rise for most 182 00:09:58,120 --> 00:10:01,880 Speaker 3: indices or corporates, particularly for weaker or second tier players. 183 00:10:02,240 --> 00:10:05,959 Speaker 3: Now that adds a vulnerability to the broader financial system. Now, 184 00:10:05,960 --> 00:10:08,040 Speaker 3: the good news is that the largest issue is are 185 00:10:08,080 --> 00:10:12,280 Speaker 3: starting from strong balance seat and substantial cash flows, and 186 00:10:12,400 --> 00:10:16,320 Speaker 3: generally with periods of technological advancement, you get higher productivity 187 00:10:16,640 --> 00:10:19,920 Speaker 3: and return on capital. So on aggregate markets too well. 188 00:10:20,320 --> 00:10:22,440 Speaker 3: But under the surface you need to be aware that 189 00:10:22,520 --> 00:10:26,240 Speaker 3: individual companies there are also clear winners and losers there. 190 00:10:26,920 --> 00:10:30,160 Speaker 1: Okay, Skyler, thanks so much for joining us. Scotta Montgomery coning. 191 00:10:30,200 --> 00:10:34,240 Speaker 1: There are Markets Live Strategist. For more explanations like this 192 00:10:34,400 --> 00:10:37,439 Speaker 1: from our team of three thousand journalists and analysts around 193 00:10:37,440 --> 00:10:42,319 Speaker 1: the world, go to Bloomberg dot com slash explainers. I'm 194 00:10:42,360 --> 00:10:45,240 Speaker 1: Stephen Carol. This is here's why. I'll be back next 195 00:10:45,280 --> 00:10:47,120 Speaker 1: week with more. Thanks for listening.