WEBVTT - AI Is Worse Than The Dot Com Bubble: Part Three 0:00:02.800 --> 0:00:07.240 Zone Media. Hello, and welcome back to Better Offline. I'm 0:00:07.240 --> 0:00:21.799 your hostead z Tron and this is the third part 0:00:21.800 --> 0:00:23.959 of our series and one way AI bubble is worse 0:00:24.000 --> 0:00:26.000 than the dot com bubble, And today we're going to 0:00:26.040 --> 0:00:29.160 focus on a question what actually burst dot com bubble. 0:00:29.720 --> 0:00:32.080 I should also be clear there is no clear answer 0:00:32.080 --> 0:00:34.519 to any question like this is kind of a kind 0:00:34.520 --> 0:00:36.920 of a difficult thing to pull apart. But to start, 0:00:36.960 --> 0:00:39.800 we're going to need a comparison, and really the InVideo 0:00:39.840 --> 0:00:42.000 of the dot com bubble would be the companies making 0:00:42.040 --> 0:00:46.120 and selling the fiber optic cables themselves and the associated hardware. 0:00:47.000 --> 0:00:50.280 In July two thousand, Corning, which made and still makes 0:00:50.320 --> 0:00:53.240 far more than just optical cables, became the largest supplier 0:00:53.320 --> 0:00:55.920 of fiber optic wires and ended two thousand and seven 0:00:55.960 --> 0:00:58.600 point one billion in revenue on an income of four 0:00:58.720 --> 0:01:01.680 hundred and twenty two million. That's profit, baby. I know 0:01:01.760 --> 0:01:03.600 we're not used to that because we talk about AI 0:01:03.720 --> 0:01:07.120 all the time. By April two thousand and one, Corning 0:01:07.160 --> 0:01:11.080 had revised its earnings estimates down three times, estimating revenue 0:01:11.080 --> 0:01:13.080 for the year to be between seven point eight billion 0:01:13.120 --> 0:01:16.720 and eight billion. Corning would eventually reveal net sales of 0:01:16.800 --> 0:01:19.800 six point two seven billion dollars with a loss of 0:01:19.880 --> 0:01:22.199 five point four ninety eight billion dollars for the year. 0:01:22.400 --> 0:01:24.840 They tried to grow a little too fast. They took 0:01:24.840 --> 0:01:27.080 out loans. They did the thing that everyone does in 0:01:27.120 --> 0:01:31.160 this kind of era. Fiber optic cable from firm JDS Uniphase, 0:01:31.200 --> 0:01:34.560 a conglomerate of different optical companies merged in nineteen ninety nine, 0:01:34.840 --> 0:01:37.680 would go on an acquisition spread to dramatically expand its 0:01:37.680 --> 0:01:40.800 fiber optic offerings, eventually crowning itself and this is from 0:01:40.800 --> 0:01:43.800 their own press release, the number one supplier of fiber 0:01:43.840 --> 0:01:46.720 ouptic components and revealing net earnings of two hundred and 0:01:46.760 --> 0:01:49.760 eight million in January two thousand and one. In July 0:01:49.800 --> 0:01:51.680 two thousand and one, it would announce a thirty five 0:01:51.680 --> 0:01:54.640 percent decrease in sales quarter over quarter and a forty 0:01:54.800 --> 0:01:57.640 four point eight billion dollar decrease in the value of 0:01:57.680 --> 0:01:59.920 companies that it had acquired, on top of a net 0:02:00.040 --> 0:02:02.200 loss of four hundred and seventy seven million dollars in 0:02:02.240 --> 0:02:04.600 the quarter. The stock would lose ninety nine percent of 0:02:04.640 --> 0:02:08.640 its value before an eventual rebrand and collapse Okay, so 0:02:08.720 --> 0:02:11.000 now you know all that crab. Let's get more specific. 0:02:11.560 --> 0:02:14.200 Nvidia currently represents seven percent of the S and P 0:02:14.320 --> 0:02:16.960 five hundred, with a market capitalization of somewhere in the 0:02:17.000 --> 0:02:21.000 region of four trillion dollars. In September two thousand, je 0:02:21.200 --> 0:02:24.480 Uniphase made up point seven percent and Colling zero point 0:02:24.560 --> 0:02:26.640 six nine two percent of the S and P five 0:02:26.720 --> 0:02:30.160 hundred when it totaled, and it is surprisingly difficult to 0:02:30.200 --> 0:02:33.639 find the exact number, around eleven point seven trillion dollars 0:02:33.639 --> 0:02:36.359 in total, with the highest weighted tech stocks being Cisco 0:02:36.400 --> 0:02:39.359 at three point one percent, Microsoft are two point five percent, 0:02:39.480 --> 0:02:41.880 Intel at two point two percent, Oracle at one point 0:02:41.919 --> 0:02:44.440 seventy five percent, and EMC at one point seven percent. 0:02:44.760 --> 0:02:46.920 Now I've had some bright sparks when I've made this 0:02:46.960 --> 0:02:51.160 point online, say ed ed, what about no hotel? Nortel 0:02:51.240 --> 0:02:53.560 was at one point four percent, Loocent was at point 0:02:53.600 --> 0:02:56.560 eight eight percent. Wasn't the same thing? In simpler terms, 0:02:56.600 --> 0:02:58.440 the dot com bubble didn't burst because of the market 0:02:58.520 --> 0:03:02.200 subsession with one company, but evenly distributed in an incredibly 0:03:02.320 --> 0:03:04.520 vague sense that the future was going to be online 0:03:04.560 --> 0:03:07.160 and we all need to be involved. The world didn't 0:03:07.200 --> 0:03:10.079 wait with baited breath for every single earnings announcement from 0:03:10.120 --> 0:03:12.959 Corning or Jduniphase, nor did the health of the market 0:03:12.960 --> 0:03:15.480 depend on the continued ability to bet and raise revenues 0:03:15.480 --> 0:03:18.720 that Echelon's higher than any other company in the market. Now, 0:03:18.760 --> 0:03:21.080 the problem with trying to find an exact comparison is 0:03:21.080 --> 0:03:25.280 that fiber optic cabling and GPUs are fundamentally different things, requiring, 0:03:25.320 --> 0:03:28.959 amongst other things, massive optical network terminals to actually turn 0:03:29.000 --> 0:03:32.560 big looms of cable into actual Internet connections. And well, 0:03:32.639 --> 0:03:34.840 I guess that you could compare that to a data center. 0:03:34.880 --> 0:03:38.200 But all of that stuff, as big as it sounds, eh, 0:03:38.240 --> 0:03:40.400 nowhere near as big as a n AI data center, 0:03:40.520 --> 0:03:43.600 Nor was it as power intensive, nor did they require 0:03:43.720 --> 0:03:46.080 to build an entire fucking gas power station in the 0:03:46.080 --> 0:03:49.480 middle of Texas for it. The massive costs associate with 0:03:49.520 --> 0:03:52.000 the fiber buildout could make Nortel or loosen to each 0:03:52.040 --> 0:03:54.680 boasted record revenues in two thousand and then got the 0:03:54.760 --> 0:03:56.680 shit kicked out of them in two thousand and one. 0:03:56.840 --> 0:03:59.839 The invidios of the dot com bubble, if you're really 0:04:00.000 --> 0:04:02.400 one though, way, I must insist that they think that 0:04:02.480 --> 0:04:05.840 we're we're going through our own thing. We're at the 0:04:05.920 --> 0:04:09.200 beginning of history, not the end of it. In any case, 0:04:09.240 --> 0:04:12.520 they're the only real revenue comparison. In Vidia's last quarterly 0:04:12.600 --> 0:04:15.080 revenue was fifty seven billion dollars, with around eighty eight 0:04:15.080 --> 0:04:17.120 percent of that coming from its state at center Vertigo, 0:04:17.240 --> 0:04:20.920 where it sells GPUs and the associated networking gear. In 0:04:21.040 --> 0:04:24.120 Loosen's case, the Service Provider Network segment, where it bundled 0:04:24.120 --> 0:04:27.839 all its telecoms infrastructure like optical networking and switching phere. 0:04:27.920 --> 0:04:30.440 It's two thousand nine yearal rapport accounted for seventy eight 0:04:30.480 --> 0:04:33.000 percent of its thirty three point eight billion dollars in revenue, 0:04:33.040 --> 0:04:35.200 a slight improvement from ninety ninety nine when it was 0:04:35.240 --> 0:04:37.360 eighty one percent and ninety ninety eight when it was 0:04:37.400 --> 0:04:40.080 eighty three percent. For Nortel, the comp was its Service 0:04:40.160 --> 0:04:42.920 Provider and Carrier segment, which made eighty two percent of 0:04:42.960 --> 0:04:45.920 its thirty billion dollars in revenue in two thousand. It 0:04:46.000 --> 0:04:47.840 gets a little confusing at this point when you try 0:04:47.880 --> 0:04:50.640 and work out whether they're either of their net incomers. 0:04:50.720 --> 0:04:53.320 Loosen had to restate revenues in two thousand and Nortel 0:04:53.360 --> 0:04:54.960 had to do so for two thousand and one, two 0:04:54.960 --> 0:04:57.080 thousand and two, and two thousand and three, thanks to 0:04:57.160 --> 0:05:01.240 creative accounting principles that, in Nortel's case, overstate revenue by 0:05:01.279 --> 0:05:05.080 nearly two billion dollars. Pobody's nerveffect I guess nevertheless, for 0:05:05.120 --> 0:05:07.440 two thousand loose and booked one point twenty one to 0:05:07.520 --> 0:05:11.000 nine billion dollars in net income, and Nortel well, okay, 0:05:11.040 --> 0:05:14.400 whoopsee doodle, they lost two billion dollars. Yes, things are 0:05:14.480 --> 0:05:17.800 materially different in the AI bubble. In Vidia posted thirty 0:05:17.839 --> 0:05:19.800 one point one to nine billion dollars in net income 0:05:19.880 --> 0:05:22.159 it's last quarter, and if we are to believe, take 0:05:22.240 --> 0:05:25.880 in Baron's marketing arm for Nvidia, who acts like a 0:05:26.000 --> 0:05:29.400 fucking cheerleader its largest customers of printing money to the 0:05:29.400 --> 0:05:32.040 point that there's simply no reason to worry. And in 0:05:32.160 --> 0:05:36.000 some ways they're right. In Vidia prints money, It's incredibly profitable, 0:05:36.240 --> 0:05:39.200 has a virtual monopoly over the GPUs behind the AI bubble, 0:05:39.400 --> 0:05:43.440 and it can effectively set prices at whatever it wants to. Similarly, 0:05:43.520 --> 0:05:46.240 the companies that are buying the most GPUs, Microsoft, Meta, 0:05:46.279 --> 0:05:49.760 Google and Amazon are all incredibly profitable themselves, which is 0:05:49.839 --> 0:05:51.560 really easy if you don't think about it for more 0:05:51.600 --> 0:05:53.520 than a minute to take to mean that in video 0:05:53.560 --> 0:05:55.800 won't end up like loosent, Nortel, or any other dot 0:05:55.880 --> 0:05:59.400 com casualty. Now, I need to be very clear about something. 0:06:00.240 --> 0:06:04.160 Just because Nvidia isn't like Nortel or loosened doesn't mean 0:06:04.240 --> 0:06:06.400 that things aren't bad. And just because some of the 0:06:06.480 --> 0:06:09.600 companies that buying these GPUs are profitable doesn't mean that 0:06:09.720 --> 0:06:12.080 all of them are or won't die on their asses 0:06:12.120 --> 0:06:16.800 the moment the debt train stopped showing up. Okay, So 0:06:16.920 --> 0:06:19.600 to really simplify the comparison here, the debt part of 0:06:19.640 --> 0:06:22.880 the dot com bubble was about customers of telecommunications companies 0:06:22.960 --> 0:06:26.720 taking on debt and the telecommunications companies themselves taking on 0:06:26.880 --> 0:06:30.080 debt to service these contracts. These companies also did a 0:06:30.160 --> 0:06:32.600 number of aggressive acquisitions, many of which had to be 0:06:32.720 --> 0:06:36.040 marked down. As I've mentioned, Loosened, Nortel and many of 0:06:36.080 --> 0:06:39.360 the other tailcom's companies building out Internet infrastructure would also 0:06:39.440 --> 0:06:42.560 loan money to their customers in a thing called vendor financing, 0:06:42.960 --> 0:06:45.159 leading them to have to write off hundreds of millions 0:06:45.200 --> 0:06:47.560 of dollars of debts. When these customers collapsed, and in 0:06:47.680 --> 0:06:50.320 fact Windstar, a company that did a massive deal with 0:06:50.520 --> 0:06:53.080 how you may remembered this with the Enron series, they 0:06:53.120 --> 0:06:56.480 ended up their bankruptcy. People ended up suing Loosen and 0:06:56.520 --> 0:06:59.560 getting hundreds of millions of dollars. In fact, that's a 0:06:59.600 --> 0:07:03.080 great place the start to make the comparison. Nortel's largest 0:07:03.120 --> 0:07:05.760 deals included a multi billion dollar contract to sell stuff 0:07:05.800 --> 0:07:09.040 to now world famous fraudster's world Com, a bizarre three 0:07:09.120 --> 0:07:12.240 billion dollar managed services deal with global IT firm computer 0:07:12.320 --> 0:07:16.080 scientists that involved two thousand Nortel employees moving into the company, 0:07:16.360 --> 0:07:18.600 and a one point four billion dollar, ten year long 0:07:18.680 --> 0:07:21.280 deal with England's Cable and Wireless Group to build and 0:07:21.400 --> 0:07:24.160 operate the cable and wireless internet backbone throughout Europe and 0:07:24.240 --> 0:07:28.120 North America that was quickly underwater. Loosen steals included a 0:07:28.160 --> 0:07:30.560 five billion dollar deal with Verizon and a one point 0:07:30.640 --> 0:07:33.400 four billion dollar deal with other carriers to supply tech 0:07:33.440 --> 0:07:36.480 to Chinese Unicorn. Money was flowing, but it was all 0:07:36.560 --> 0:07:39.440 in these weird directions with people that couldn't necessarily afford 0:07:39.480 --> 0:07:56.200 it or that would lose. Everybody involve money. The comparison 0:07:56.240 --> 0:07:59.840 starts to drift when you talk about revenue centralization. Loosened, 0:08:00.080 --> 0:08:03.080 Nortel and other dot com bubble telecoms companies had a 0:08:03.200 --> 0:08:05.520 variety of deals in the one hundred million dollars to 0:08:05.600 --> 0:08:07.960 three hundred million dollar range and many more that were 0:08:08.000 --> 0:08:10.440 in the twenty million dollars to fifty million dollar range. 0:08:10.600 --> 0:08:12.960 In two thousand eight, and t account for ten percent 0:08:13.000 --> 0:08:16.960 of Loosen's revenue and Verizon thirteen percent. Nortel sadly didn't 0:08:17.000 --> 0:08:21.280 disclose its revenue breakdown by customer. Nortel was, however, Calling's 0:08:21.360 --> 0:08:24.520 largest customer for optical cable and fun fact, Loosen also 0:08:24.560 --> 0:08:26.920 made optical cable and was the second largest provider in 0:08:26.960 --> 0:08:30.880 the business. Despite what JDA Uniphase said, JDA uniforse, you 0:08:31.040 --> 0:08:34.280 fucking lied to me that one's just for Phil. Phil 0:08:34.360 --> 0:08:36.120 brought on. You're listening to this, you're going to hear 0:08:36.200 --> 0:08:40.400 that you can go. I fucking hated jediuniphase or JDS uniphase. 0:08:40.559 --> 0:08:43.400 I learned about this company a month ago. I hate them, 0:08:44.280 --> 0:08:47.680 but that revenue centralization is very important. In Video's revenue 0:08:47.760 --> 0:08:51.559 is extremely centralized. In its last earnings, in Video noted 0:08:51.600 --> 0:08:54.319 that sixty one percent of its revenue came from four 0:08:54.440 --> 0:08:57.400 unnamed companies, and for years somewhere between eighteen percent and 0:08:57.520 --> 0:08:59.800 forty percent of its revenue has come from anywhere from 0:09:00.000 --> 0:09:03.079 doing two and four companies, none of which it names. 0:09:04.320 --> 0:09:07.360 This is critically important because in video represents, as I said, 0:09:07.440 --> 0:09:09.199 seven percent of the value of the S and B 0:09:09.320 --> 0:09:12.480 five hundred and the markets have an unhealthy relationship with 0:09:12.559 --> 0:09:15.560 its stock freaking out in August twenty twenty five, when 0:09:15.640 --> 0:09:17.559 year of a year growth was only I am not 0:09:17.760 --> 0:09:20.400 fucking with you, expected to be around fifty percent year 0:09:20.480 --> 0:09:24.920 every year, people were freaking out. They dumped the stock 0:09:24.960 --> 0:09:28.040 for over a month. It was crazy. As a result, 0:09:28.120 --> 0:09:30.719 any stock panic caused by Nvidia will naturally drag down 0:09:30.800 --> 0:09:34.079 the entire market with him. The dot com bubble was, 0:09:34.200 --> 0:09:37.000 as I've discussed two things, the bullshit dot com bubble 0:09:37.360 --> 0:09:40.840 from websites and the telecommunications bubble caused by massive overbuilds 0:09:40.880 --> 0:09:44.160 of fiber optic and internet services. When the bubble burst, 0:09:44.200 --> 0:09:46.600 it was caused by rising interest rates increasing the cost 0:09:46.679 --> 0:09:49.640 of borrowing, and the markets realizing that these unprofitable companies 0:09:49.679 --> 0:09:53.240 wouldn't survive long term, and then venture capital being depleted 0:09:53.559 --> 0:09:55.160 and I realized it's tempting to claim we're in the 0:09:55.200 --> 0:09:58.360 same situation, but I must I must insist it's really 0:09:58.440 --> 0:10:01.959 quite different. At the time, venture capital was much much smaller. 0:10:02.280 --> 0:10:03.880 For the following numbers, I'm going to give you the 0:10:03.960 --> 0:10:05.959 numbers sug justed for inflation, otherwise I'm going to be 0:10:06.000 --> 0:10:10.040 here all fucking day. US venture capital invested twenty three 0:10:10.120 --> 0:10:12.960 billion dollars in ninety ninety seven, twenty eight point twenty 0:10:13.000 --> 0:10:15.440 one billion dollars in ninety ninety eight, ninety five point 0:10:15.480 --> 0:10:17.920 five billion dollars in ninety ninety nine, and one hundred 0:10:17.920 --> 0:10:20.440 and ninety seven point seven one billion dollars in two thousand, 0:10:20.559 --> 0:10:22.439 for a grand total of three hundred and forty four 0:10:22.520 --> 0:10:25.559 point five billion dollars, a mere six point two billion 0:10:25.600 --> 0:10:28.000 dollars more than the three hundred and thirty eight point 0:10:28.120 --> 0:10:30.559 three billion dollars raised in twenty twenty five alone in 0:10:30.640 --> 0:10:33.360 venture capital was somewhere between forty and fifty percent of 0:10:33.400 --> 0:10:35.800 that around one hundred and sixty eight billion dollars going 0:10:35.800 --> 0:10:38.840 into AI investments, and in twenty twenty four, North American 0:10:38.880 --> 0:10:42.240 AI startups only raised around one hundred and six billion dollars. 0:10:43.040 --> 0:10:46.640 HiT's growing HiT's happening more the dot com bubble burst 0:10:46.679 --> 0:10:48.760 when the bullshit dot com stocks died on their ass 0:10:48.880 --> 0:10:50.680 and the world realized that the magic of the Internet 0:10:50.800 --> 0:10:53.199 was not a panacea that would fix every business model, 0:10:53.320 --> 0:10:55.199 and there was no magic moment where a company like 0:10:55.280 --> 0:10:59.079 webvan or pets dot Com would turn into this magical 0:10:59.240 --> 0:11:03.280 profitable beast from a horribly unprofitable business. Similarly, companies like 0:11:03.360 --> 0:11:06.319 Loosent Technology stop being rewarded for doing dodgy circuit that 0:11:06.360 --> 0:11:09.000 deals with companies like Windstar, leading to the collapse of 0:11:09.040 --> 0:11:11.760 the telecommunications bubble that led to millions of miles of 0:11:11.880 --> 0:11:14.839 dark fiber being sold dirt cheap in two thousand and two. 0:11:15.520 --> 0:11:18.200 The oversupply of dark fiber was eventually seen as a positive, 0:11:18.280 --> 0:11:20.800 leading to an eventual surge and demand as billions of 0:11:20.800 --> 0:11:24.559 people came online toward the end of the two thousands. Now, 0:11:24.640 --> 0:11:27.320 I know what you're thinking, Ed. Isn't this exactly what's 0:11:27.360 --> 0:11:30.600 happening here in the AI bubble? Isn't this the same thing. 0:11:30.880 --> 0:11:34.560 We've got overvalued stars, We've got multiple unprofitable, unsustainable AI 0:11:34.679 --> 0:11:38.079 companies promising to IPO, We've got overvalued tech stocks, and 0:11:38.120 --> 0:11:40.960 we've got one of the largest infrastructure buildouts of all time. 0:11:41.280 --> 0:11:43.800 Tech companies are trading at ridiculous multiples of their earnings 0:11:43.840 --> 0:11:47.719 per share, but the multiples aren't as high. That's good, right, No, 0:11:47.960 --> 0:11:48.360 it isn't. 0:11:48.640 --> 0:11:49.040 It isn't. 0:11:49.280 --> 0:11:51.719 AI boosters and well wishers are obsessed with making this 0:11:51.800 --> 0:11:54.800 comparison because saying things worked out after the dot com 0:11:54.880 --> 0:11:59.120 bubble allows them to rationalize doing stupid, destructive, and shitty things. 0:12:00.040 --> 0:12:02.280 Even if this was just like the dot com bubble, 0:12:02.480 --> 0:12:07.200 things would be absolutely fucking catastrophic. The NASDAG drops seventy 0:12:07.240 --> 0:12:09.959 eight percent from its peak in March two thousand, this 0:12:10.200 --> 0:12:12.560 time due to the incredible ignorance of both the private 0:12:12.600 --> 0:12:16.120 and public powerbrocus of the tech industry. I expect consequences 0:12:16.559 --> 0:12:20.880 that will range from horrifying to calamitus, depending almost entirely 0:12:20.920 --> 0:12:22.959 on how long the bubble takes to burst and how 0:12:23.000 --> 0:12:26.280 willing the SEC is the greenlight and IPO of Open 0:12:26.360 --> 0:12:30.480 AI or anthropic. I'm very worried, and tomorrow I'm going 0:12:30.520 --> 0:12:32.839 to finish up this series with a somewhat dark note. 0:12:33.200 --> 0:12:34.880 Then we need to stop pretending that this will be 0:12:34.920 --> 0:12:37.400 a smooth landing or that anything will be left in 0:12:37.440 --> 0:12:38.600 its wake worth keeping. 0:12:47.240 --> 0:12:49.640 Thank you for listening to Better Offline. The editor and 0:12:49.679 --> 0:12:52.839 composer of the Better Offline theme song is Matasowski. You 0:12:52.880 --> 0:12:55.080 can check out more of his music and audio projects 0:12:55.280 --> 0:13:01.520 at Mattasowski dot com m attoso wa Ski dot com. 0:13:02.240 --> 0:13:04.520 You can email me at easy at Better offline dot 0:13:04.600 --> 0:13:06.800 com or visit Better Offline dot com to find more 0:13:06.840 --> 0:13:10.200 podcast links and of course, my newsletter. I also really 0:13:10.240 --> 0:13:12.480 recommend you go to chat dot where's youreaed dot at 0:13:12.559 --> 0:13:15.000 to visit the discord, and go to our slash Better 0:13:15.040 --> 0:13:18.160 Offline to check out our reddit. Thank you so much 0:13:18.240 --> 0:13:22.040 for listening. Better Offline is a production of cool Zone Media. 0:13:22.200 --> 0:13:25.040 For more from cool Zone Media, visit our website cool 0:13:25.120 --> 0:13:28.199 Zonemedia dot com, or check us out on the iHeartRadio 0:13:28.240 --> 0:13:28.440 app 0:13:28.480 --> 0:13:30.920 Apple Podcasts, or wherever you get your podcasts.