YouTube2Text

How AI is Draining Google’s Finances — Transcript

by TLDR Business · 1,508 words · 244 segments · language en · Watch on YouTube

Full transcript

  1. 0:00On Wednesday, Google's parent company
  2. 0:01Alphabet published its financial results
  3. 0:03for the third quarter of this year, and
  4. 0:05well, they weren't well received.
  5. 0:08Spending came in hot, revenue came in
  6. 0:10soft, and the stock duly slumped by 10%
  7. 0:13in after-hours trading, wiping about
  8. 0:15$400 off Alphabet's market cap in barely
  9. 0:19a few hours. So, in this video, we're
  10. 0:21going to explain why Alphabet stock just
  11. 0:23slumped, and why, shock, it's all to do
  12. 0:25with AI.
  13. 0:33Sick of slop? Well, Tail Digger's
  14. 0:35content is entirely AI-free, with us
  15. 0:37having strict rules when it comes to the
  16. 0:39use of AI in our company. So, while we
  17. 0:41might not be perfect, at least you know
  18. 0:44we're human. Find our full AI policy
  19. 0:46linked in the description.
  20. 0:48So, simply put, the fundamental problem
  21. 0:50for Google is that spending came in
  22. 0:52higher than expected, and revenues came
  23. 0:54in lower than expected. And well, you
  24. 0:56don't need an MBA to know that's not
  25. 0:58good business. Let's start with
  26. 1:00spending. In the third quarter alone,
  27. 1:02which is July to September this year,
  28. 1:04Google spent a whopping $87 billion,
  29. 1:07which is up 22% from the same period in
  30. 1:102025. This follows a broad trend, with
  31. 1:13spending in the year-to-date up 21%
  32. 1:16reaching $236 billion in the first three
  33. 1:19quarters. Obviously, the culprit here is
  34. 1:21AI, but what does that mean? Well, we
  35. 1:24can break down Google's soaring AI spend
  36. 1:26into two main categories: capital
  37. 1:28expenditure and TPU inventory build-up.
  38. 1:31If those things mean nothing to you,
  39. 1:33don't worry, we'll explain each in
  40. 1:34detail. And when we're done, you should
  41. 1:36have a clear idea of what hyperscalers
  42. 1:38are actually spending money on as they
  43. 1:40fund the AI build-outs. First, capital
  44. 1:43expenditure, or capex. This refers to
  45. 1:45money being spent up front by Google
  46. 1:47towards its own infrastructure. As you
  47. 1:50can see on their balance sheet, this
  48. 1:51value of the company's physical assets
  49. 1:54has leapt up since last year, which
  50. 1:56reflects growing capex investments.
  51. 1:58Right there, you can see $51.4 billion
  52. 2:01spent in the last quarter alone, which
  53. 2:03we're told is funding Google's AI build
  54. 2:05out. But, what on earth is that money
  55. 2:07being spent on? Well, according to CFO
  56. 2:10Anat Ashkenazi, it's a 65/35%
  57. 2:13split between active computing hardware
  58. 2:15and physical data center infrastructure.
  59. 2:18The computing stuff is essentially GPUs,
  60. 2:20CPUs, and TPUs, along with the tech that
  61. 2:23allows huge chip clusters to communicate
  62. 2:25with each other. The data centers costs
  63. 2:27go towards physical buildings, their
  64. 2:29energy supplies, and cooling systems.
  65. 2:32So, that's CapEx. Next is TPU inventory.
  66. 2:35Google's AI build out strategy is
  67. 2:37two-pronged. On the one hand, they're
  68. 2:39building out their own computing power
  69. 2:40so they can reap the rewards of the AI
  70. 2:42boom. On the other hand, they're acting
  71. 2:44a bit like Nvidia and designing their
  72. 2:47own specialized AI chips called TPUs, or
  73. 2:50tensor processing units, which they plan
  74. 2:52to sell on to clients like Anthropic,
  75. 2:54Apple, and Oracle. If you look at that
  76. 2:56inventory line on the balance sheet, you
  77. 2:58can see that since last year, an
  78. 3:00additional $10.9 billion
  79. 3:02has been spent building up Google's
  80. 3:04inventory. A few years ago, that line
  81. 3:06would have reflected products like Pixel
  82. 3:07phones sitting in warehouses waiting to
  83. 3:10be sold. Now, we're told that the figure
  84. 3:12is almost entirely accounted for by TPU
  85. 3:14building. But, before you leap to
  86. 3:16conclusions, there's an important point
  87. 3:18here to keep in mind, which illustrates
  88. 3:20that Google's AI spend isn't without
  89. 3:22justification. Google's Cloud CEO,
  90. 3:24Thomas Kurian, pointed out in
  91. 3:26Wednesday's earnings call that the
  92. 3:27company only produces components and
  93. 3:29builds TPU inventory after signing
  94. 3:32contracts with customers. In other
  95. 3:33words, while this multi-billion-dollar
  96. 3:35figure is sticking out on their books
  97. 3:37like a sore thumb, those processors are
  98. 3:39going to be shipped soon and turn into
  99. 3:41revenue.
  100. 3:42In broad terms, this speaks to a dynamic
  101. 3:44in the AI build out which benefits
  102. 3:46hyperscalers like Google, Amazon, and
  103. 3:48Microsoft over companies further down
  104. 3:50the food chain like OpenAI and
  105. 3:52Anthropic. Sure, the hyperscalers are
  106. 3:54spending huge amounts building out
  107. 3:56America's computing power, but they also
  108. 3:58have significant commitments from
  109. 4:00enterprise customers, which, if all goes
  110. 4:02well, means huge revenues will follow.
  111. 4:04Of course, we'll be the first to admit
  112. 4:06that if all goes well is doing a lot of
  113. 4:08heavy lifting in that sentence. That
  114. 4:10brings us to the other side of the
  115. 4:12ledger, revenues. At first glance,
  116. 4:14revenues look pretty good. Total revenue
  117. 4:17for the quarter came in at $127.6
  118. 4:20billion,
  119. 4:21only a tenth below analyst forecast of
  120. 4:23$128 billion,
  121. 4:25and up 25% compared to a year ago, i.e.
  122. 4:29to the third quarter 2025. The thing
  123. 4:31that slightly spooked the markets,
  124. 4:33however, was where that revenue came
  125. 4:35from. For context, Alphabet's revenues
  126. 4:37are often broken down into three
  127. 4:39segments. Google services, which covers
  128. 4:41stuff like ad revenues from Google
  129. 4:43Search and YouTube, as well as
  130. 4:45subscription fees, Google Cloud, which
  131. 4:47basically rents out computing power and
  132. 4:49software to businesses and other
  133. 4:50organizations, so they don't have to
  134. 4:52build and run their own data centers,
  135. 4:54and other bets, which covers the
  136. 4:57loss-making more experimental
  137. 4:58businesses, and is usually a relatively
  138. 5:00small slice of the overall pie.
  139. 5:03Historically, Google services has
  140. 5:04accounted for the vast majority of all
  141. 5:06revenues, and has been by far and away
  142. 5:09the most profitable part of the
  143. 5:10business. The thing that spooked the
  144. 5:12markets this time round is that Google
  145. 5:14services revenue came in below forecast
  146. 5:17at $96 billion.
  147. 5:19That's up 11% year-on-year, but down
  148. 5:21slightly compared to the previous
  149. 5:23quarter. This was in part because
  150. 5:24revenues from Google Search,
  151. 5:26historically by far the largest and most
  152. 5:28profitable part of Alphabet's business,
  153. 5:30came in at $63.3 billion, about $1.5
  154. 5:34billion below forecast, and basically
  155. 5:36flat on the previous quarter. The only
  156. 5:38reason that headline revenues came in
  157. 5:40line with expectations is because of an
  158. 5:42above-forecast print from Google Cloud,
  159. 5:44which generated about $30 billion in
  160. 5:46revenues for the quarter, roughly $2
  161. 5:49billion more than forecast. Now, at this
  162. 5:51point you might be asking, "What's wrong
  163. 5:53with all this?" It sounds like Alphabet
  164. 5:55is essentially diversifying, becoming
  165. 5:57less singularly reliant on the services
  166. 5:59part of the business, and offsetting any
  167. 6:01shortfalls over there with more revenues
  168. 6:04elsewhere, like Google Cloud. And yet,
  169. 6:06while diversification is generally a
  170. 6:08good thing, Google services has been a
  171. 6:10supremely reliable money printing
  172. 6:12machine for Alphabet for years now. And
  173. 6:15markets apparently aren't convinced that
  174. 6:17other parts of the business, like AI,
  175. 6:19can truly fill its shoes.
  176. 6:21Google Search, for instance, has been a
  177. 6:23money printing machine in large part
  178. 6:25because it's had a near monopoly on the
  179. 6:27search engine market for years, if not
  180. 6:29decades now. It's nice that Google Cloud
  181. 6:32is making money, but it only became
  182. 6:34consistently profitable in around 2023.
  183. 6:37And it's not obvious, at least to the
  184. 6:38markets, why it would ultimately be as
  185. 6:41profitable as, say, Google Search, given
  186. 6:43the lack of any obvious moat.
  187. 6:46At the same time, the trend lines look
  188. 6:47like they're squarely headed in the
  189. 6:49wrong direction for search at the
  190. 6:50moment, as people move away from
  191. 6:52traditional search towards AI chatbots.
  192. 6:55Alphabet themselves admitted on
  193. 6:57Wednesday that the growing popularity of
  194. 6:59Google's AI mode was taking a toll on
  195. 7:01search revenues, because AI queries
  196. 7:03apparently have lower monetization
  197. 7:05rates.
  198. 7:06All in all, this combination of lower
  199. 7:07revenues and lots more spending meant
  200. 7:09that Alphabet's free cash flow came in
  201. 7:12at negative $9 billion, down from
  202. 7:14negative $5.9 billion in the previous
  203. 7:17quarter, and only the second quarter in
  204. 7:19the red since Alphabet went public in
  205. 7:212004. So, that's Alphabet's financial
  206. 7:23situation right now. They're in a
  207. 7:25somewhat risky position, which explains
  208. 7:27the stock dip this week. But there's
  209. 7:29certainly room to recover if their bet
  210. 7:32on AI pays off. Exciting announcement.
  211. 7:35I'm happy to reveal the cover of the
  212. 7:36next issue of Two Long, our 10th. The
  213. 7:39main story is all about the ascendancy
  214. 7:41of China, spanning energy, artificial
  215. 7:43intelligence, global trade, and an
  216. 7:45increasingly assertive posture towards
  217. 7:47its neighbors. The country's trajectory
  218. 7:49points towards total dominance. Yet
  219. 7:52beneath the surface of this meteoric
  220. 7:53rise, cracks may be forming. You can
  221. 7:55pre-order your copy right now, and if
  222. 7:57you use code 10, you'll get £3 off any
  223. 8:00new subscription, including the Two Long
  224. 8:02and TLDR party bundle subscription. Also
  225. 8:05also mentioning, as always, there's much
  226. 8:07more than just the cover story. In fact,
  227. 8:09and I know you'll love it when I say
  228. 8:11this, your talk of our longest issue
  229. 8:13ever is so repetitive.
  230. 8:15>> I don't want to hear about your longest
  231. 8:16issue ever.
  232. 8:17>> It's only one or two pages more than the
  233. 8:19other issues anyway. But this time it
  234. 8:21truly is longer at 100 pages to
  235. 8:24celebrate our 10th issue. That means
  236. 8:26that we also have tons of space to
  237. 8:27discuss the US midterms, the return of
  238. 8:29company towns, the future of Taiwan, the
  239. 8:32seeming rise of the AFD, Burnham
  240. 8:34reversing Thatcherism, Ukraine's
  241. 8:35corruption scandals, and much more. Find
  242. 8:38out more about the latest issue and
  243. 8:39subscribe to Too Long by heading to our
  244. 8:41website. It's linked in the description.

About this transcript

This page contains the full transcript of How AI is Draining Google’s Finances by TLDR Business, generated from the public captions YouTube serves with the video. The transcript has 1,508 words across 244 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.