SpaceX Went Public - A Disaster Waiting to Happen — Transcript

Full transcript

1. 0:00If you give me $500, would you invest
2. 0:01into SpaceX? I I I think I would, yeah.
3. 0:03>> No, I wouldn't that.
4. 0:04>> You would not.
5. 0:05>> No, but I hope my index funds aren't
6. 0:07automatically doing it.
7. 0:07>> Hi, welcome to another episode of
8. 0:09ColdFusion. SpaceX is an incredible
9. 0:12company with impressive achievements. It
10. 0:14brought the price of space launches down
11. 0:16by 85%. Objectively, that is
12. 0:18astonishing. The company also made
13. 0:21reusable rockets reality. Many critics
14. 0:23said that was impossible. If you love
15. 0:24your engineering and physics, yep, great
16. 0:26[music] company. But now, there's a
17. 0:28massive problem. Recently, something has
18. 0:31gone terribly wrong with the company.
19. 0:32It's no longer what most people think it
20. 0:34is. Let me ask you a question. What type
21. 0:37of company do you think SpaceX is? You
22. 0:39might say a rocket cargo company that
23. 0:41operates satellite communications. For
24. 0:43those who are more optimistic, you could
25. 0:45say that SpaceX is an exploration
26. 0:47company who will take humanity to the
27. 0:49moon and beyond. Well, according to
28. 0:51their own S-1 filing, the documents a
29. 0:53company needs to file before it goes
30. 0:54public on the stock market, both of
31. 0:56those answers are wrong.
32. 0:58Generally, S-1 filings list the industry
33. 1:00category that a company is under. For
34. 1:02example, the industry code 3760 is for
35. 1:05spacecraft, 3182 is for defense and
36. 1:08aeronautical. Other companies like
37. 1:10Virgin Galactic, Boeing, etc., all file
38. 1:13under these categories. But for SpaceX,
39. 1:15they filed under the industry code of
40. 1:177370.
41. 1:19That's computer programming and data
42. 1:21processing. Oh, it looks like SpaceX is
43. 1:24officially filing to be an AI data
44. 1:26center company. That already raises an
45. 1:28eyebrow, but it's worse than that. To
46. 1:30understand how, you first need to
47. 1:32understand what a TAM is. TAM stands for
48. 1:34total addressable market. It's the
49. 1:36maximum revenue opportunity available
50. 1:38for a product or service if you achieved
51. 1:40100% market share. It basically
52. 1:42represents to your investors what the
53. 1:44market for your product is.
54. 1:46Looking at the prospectus, SpaceX's
55. 1:48[music] TAM is 28.5 trillion,
56. 1:50approximately the size of the GDP of the
57. 1:52United States. Okay, that's pretty high,
58. 1:54but that's not the bizarre part. The
59. 1:55strangest part is only 15% of this
60. 1:58market is space and communications, and
61. 2:00it's written right here in front of your
62. 2:02face in plain text on page 11.
63. 2:05So, 85% of the company is essentially
64. 2:08just a wrapper for yet another LLM and
65. 2:11data center business [music]
66. 2:12like Anthropic or OpenAI. I know some of
67. 2:15you guys have commented [music] that
68. 2:16you're getting sick of hearing about AI,
69. 2:18but this isn't funny anymore. It's
70. 2:20getting really serious.
71. 2:21When SpaceX went [music] public, it was
72. 2:23the largest launch of any company in the
73. 2:26history of financial markets, [music]
74. 2:28larger than Amazon, Google, Alibaba, and
75. 2:31even Saudi Aramco.
76. 2:33In the first few hours of trading,
77. 2:35SpaceX [music] blasted past a valuation
78. 2:37of $2.3 trillion,
79. 2:39instantly becoming the seventh most
80. 2:41valuable company and worth approximately
81. 2:43the same as the country of Canada.
82. 2:45For some people, this seemed like a sign
83. 2:47of strength, but for others, it was
84. 2:48terrifying, the signs of an overheated
85. 2:51market. But, that's not all.
86. 2:54The story of the SpaceX IPO is telling
87. 2:57us something if we listen closely. It
88. 2:58shines a light on the state of the
89. 3:00current financial markets. A corporation
90. 3:03is hiding an AI company losing a billion
91. 3:05dollars a month inside [music] the brand
92. 3:06of a perfectly healthy rocket company
93. 3:08just so investors fund it. Let me say
94. 3:11that again. In 2026, a corporation can
95. 3:14disguise an AI company losing almost a
96. 3:16billion dollars a month inside of a
97. 3:17perfectly healthy rocket company, sell
98. 3:19it to investors, and get away with it.
99. 3:22>> You know, your expectation is not going
100. 3:23to be a good investment in that
101. 3:25>> That's the part that I'm trying to
102. 3:26understand. I found that that it just
103. 3:28it's never
104. 3:29been a great bet to bet against Elon.
105. 3:32>> I I'll I'll [music] grant you all that.
106. 3:34>> If you again, look at the S1, the TAM,
107. 3:36the total addressable market is defined
108. 3:38at 28.5 trillion.
109. 3:4185% of that TAM
110. 3:44>> [music]
111. 3:44>> is in AI. So, even if everything you say
112. 3:47about SpaceX and Starlink [music] is
113. 3:49true, the entire company is being bet on
114. 3:53AI in in terms of of its future, not on
115. 3:56space and and not on Starlink.
116. 4:00Those are his words, not mine.
117. 4:02>> It's a mess, and in this episode, it's
118. 4:04going [music] to be a wild ride. And a
119. 4:05disclaimer, none of this is financial
120. 4:07advice. If you want to buy the stock,
121. 4:10buy it. If you don't, don't. But let's
122. 4:12get into it.
123. 4:15>> You are watching ColdFusion TV.
124. 4:18>> There's a lot of excitement about what's
125. 4:20going to happen.
126. 4:20>> Everything they can possibly do to
127. 4:22guarantee a pop is being done.
128. 4:24>> This is going to become the largest
129. 4:26company on the planet.
130. 4:28>> For over two decades, SpaceX earned its
131. 4:30dominance the hard and old-fashioned
132. 4:32way. Three catastrophic early failures,
133. 4:35a near bankruptcy that Musk described as
134. 4:37existential, a methodically driven slow
135. 4:39rise, and finally, a near monopoly on
136. 4:42orbital launches.
137. 4:43In a company full of moonshots, if we
138. 4:45isolate Starlink, it's the only thing
139. 4:47that actually makes money. The Starlink
140. 4:49internet service has a constellation of
141. 4:51over 9,000 satellites [music] and
142. 4:52generated 11.4 billion in revenue and
143. 4:554.4 billion in operating profit in 2025.
144. 4:58But the rest of the company is another
145. 5:00story, and trust me, we'll get to that.
146. 5:02And from here, the controversy begins.
147. 5:08When people talk about the broader
148. 5:09market, they're referring to the S&P
149. 5:11500. It includes all the big names like
150. 5:13Apple, Nvidia, Microsoft, and Amazon.
151. 5:16SpaceX had hoped to fast-track its way
152. 5:18into the S&P 500 immediately upon
153. 5:21listing. This was essentially changing
154. 5:22the rules. They were essentially forcing
155. 5:25the general public to buy the stock.
156. 5:27Fast-tracking its way into the index
157. 5:28would have forced every 401k and even
158. 5:30pension funds to buy the stock. This
159. 5:32isn't just American citizens we're
160. 5:34talking about, but investment funds of
161. 5:36everyday people all around the world.
162. 5:38Roughly 14 billion of automatic passive
163. 5:41fund inflows going straight [music] into
164. 5:42SpaceX. It would have been a massive
165. 5:44wealth transfer, but thankfully, the
166. 5:46index committee said no.
167. 5:48Why?
168. 5:49SpaceX was just too risky. To join the
169. 5:52S&P, a company first must be profitable
170. 5:54in its most recent four quarters. SpaceX
171. 5:57didn't meet the mark. Despite posting 18
172. 5:59billion in revenue in 2025, which was a
173. 6:0233% year-over-year increase, it lost 5
174. 6:05billion for the same [music] year. So,
175. 6:07the S&P 500 refused to play ball, and
176. 6:09now that $14 billion in passive funds is
177. 6:11off the table. But, the Nasdaq, [music]
178. 6:13they said yes. So, there still will be
179. 6:15some passive funding.
180. 6:16>> And because SpaceX will be fast-tracked
181. 6:18onto the Nasdaq index, major funds will
182. 6:21be forced to buy some shares.
183. 6:23>> A lot of Australian investors will be
184. 6:26exposed to SpaceX whether they want to
185. 6:28be or not.
186. 6:32>> SpaceX spent years dismissing the public
187. 6:34markets. Elon Musk had watched what
188. 6:36quarterly earnings obsession did to
189. 6:38Tesla.
190. 6:39For a company that measures success in
191. 6:41orbital mechanics and decades,
192. 6:43impressing shareholders in a 90-day
193. 6:44cycle seemed like a poor fit.
194. 6:47In fiscal year 2023, SpaceX's CapEx was
195. 6:50manageable at 42% of revenue. And in
196. 6:532024, SpaceX was even profitable,
197. 6:56clearing about 791 million in net
198. 6:58income. For an expense-heavy rocket
199. 7:00company, that is remarkable. So, what
200. 7:03changed? Two letters, AI. In February of
201. 7:072026, SpaceX merged with xAI. Now, the
202. 7:11company's performance was in shambles.
203. 7:13>> [music]
204. 7:13>> After the merge, the required
205. 7:15retrospective accounting attributed to
206. 7:17that net loss of 5 billion we talked
207. 7:18about in 2025. xAI is the company that's
208. 7:21bleeding so much money that it needed
209. 7:23the public to fund it. Starlink [music]
210. 7:25profits alone weren't nearly enough.
211. 7:27>> Career market analysts like Michael
212. 7:29McCarthy say investors should be
213. 7:30extremely cautious. [music]
214. 7:32>> No matter how we cut the numbers, they
215. 7:34simply don't add up.
216. 7:36>> In 2025, xAI was losing $2 for every
217. 7:39dollar earned,
218. 7:40>> [clears throat]
219. 7:40>> burning through approximately $28
220. 7:42million per day, quarterly losses
221. 7:44accelerated through the year. By Q1 of
222. 7:462026, XAI's capital expenditure alone
223. 7:50reached $7.7 billion.
224. 7:52That's an annualized pace of 30.8
225. 7:54billion, more than double the entire
226. 7:56prior year. In that same period,
227. 7:58SpaceX's capital expenditure had gone
228. 8:00from that manageable 42% of revenue to
229. 8:03215% of revenue.
230. 8:06Objectively, SpaceX [music]
231. 8:07was a great company by itself. There was
232. 8:09no need to shove AI into it. Every
233. 8:12dollar that XAI burns is a dollar that
234. 8:14the rest of the company could have used.
235. 8:16>> [music]
236. 8:16>> So, to summarize, in one transaction, a
237. 8:19focused aerospace business became a
238. 8:21conglomerate of rockets, satellites,
239. 8:23[music]
240. 8:23social media, and a chatbot in the form
241. 8:26of Grok. Up next, we'll see some of the
242. 8:28bizarre things happening with XAI, and
243. 8:30then we'll take a close look at SpaceX's
244. 8:32biggest bet, data centers in space.
245. 8:35But first, this message is sponsored by
246. 8:38Raycon. Finding a gift that's actually
247. 8:40useful is the hard part, but Raycon's
248. 8:42essential open ear buds are one of those
249. 8:44things I can genuinely see getting used
250. 8:45every day. [music] I've been using them
251. 8:47quite a bit recently on walks, while
252. 8:49running errands, and even working from
253. 8:50home. What makes them different is the
254. 8:52open ear design. They [music] sit just
255. 8:54outside your ear canal, so you can
256. 8:55listen to music, podcasts, or calls
257. 8:58while still hearing what's going on
258. 8:59around you. Whether I'm walking through
259. 9:01the neighborhood or just getting outside
260. 9:03after working on a script the whole day,
261. 9:04I don't feel completely disconnected
262. 9:06from everything around me. They're also
263. 9:08incredibly lightweight, and the flexible
264. 9:10multi-angler ear hook means that they
265. 9:12stay secure without becoming
266. 9:13uncomfortable. Whether I've worn them on
267. 9:15long walks, around the house, or during
268. 9:17workouts, after a while, you honestly
269. 9:19forget you're wearing them. A few
270. 9:20features that make this a go-to pair of
271. 9:22headphones is the open ear design, the
272. 9:24comfortable fit, and the multi-point
273. 9:25connection. I can switch between my
274. 9:27phone and laptop without having to
275. 9:29reconnect anything, and the battery life
276. 9:30is also excellent, up to 36 hours with
277. 9:33the charging case. I charge them maybe
278. 9:35once a week. Raycon now has over 3
279. 9:37million customers [music] and the sound
280. 9:39quality is right up there with the far
281. 9:40more expensive brands. Plus, they come
282. 9:42with a 30-day guarantee, so there's
283. 9:44really no risk trying them out. I got
284. 9:46one of these for Mother's Day, too, and
285. 9:48she started using it as well. So, if
286. 9:49you're looking for something practical
287. 9:51this Father's Day, this is a great
288. 9:52option. Upgrade your dad's [music]
289. 9:54workout routine. Go to
290. 9:55buyraycon.com/coldfusionopen
291. 9:58to get 15% off. Thanks to Raycon for
292. 10:00supporting Cold Fusion. Now, let's get
293. 10:02back to these data centers in space.
294. 10:07Since AI is 85% of the company's
295. 10:09proposed market, we're forced to look at
296. 10:11how xAI is performing in the LLM
297. 10:13landscape.
298. 10:15So, users switch between ChatGPT,
299. 10:17Claude, Gemini, and Grok consistently.
300. 10:20Critics argue that AI is essentially a
301. 10:23commodity at this point. Nobody talks
302. 10:25about Grok being the market leader
303. 10:27except for users on the X platform. In
304. 10:29fact, according to some estimates, Grok
305. 10:31only has 0.4%
306. 10:33of enterprise use. In the AI sector,
307. 10:36enterprise is where all the money is.
308. 10:37So, that's not a good sign. All right,
309. 10:39well, maybe the management team is a
310. 10:41star-studded lineup of smart people, so
311. 10:43they'll figure it out, right?
312. 10:44Well, perhaps not. Musk founded xAI with
313. 10:4711 co-founders in 2023. Today, all 11
314. 10:51have left. There's zero remaining.
315. 10:54Okay, so that's not encouraging, but
316. 10:56that's still fine. They can still make
317. 10:58money by selling compute to other AI
318. 10:59companies. Well, there's some problems
319. 11:01there, too. The deals so far don't give
320. 11:03an aura of confidence.
321. 11:06Google entered a partnership with xAI to
322. 11:08lease some GPU compute for $920 million
323. 11:10a month. Sounds great, but when you look
324. 11:13deeper into the deal, there's a strange
325. 11:15smell. Firstly, either party can
326. 11:17terminate the contract with 90 days
327. 11:19notice. Okay, that's a bit tenuous, but
328. 11:21reading further, you learn that Google
329. 11:23has a 6% stake in xAI, and there you go.
330. 11:27That smell gets stronger. It smells like
331. 11:29yet more circular financing with a
332. 11:31sprinkle of revenue boosting for the
333. 11:33SpaceX IPO season. Okay, so what about
334. 11:35Anthropic? Didn't they sign a deal with
335. 11:37SpaceX in May of 2026? Yes, but the way
336. 11:40that deal came about is hilarious. It
337. 11:42turns out that XAI screwed up the design
338. 11:45for their first mega data center. For
339. 11:46some reason, they built it with three
340. 11:48different types of GPUs. This meant that
341. 11:51the fastest GPUs, which are sitting
342. 11:52there idle, waiting for the slower ones
343. 11:54to finish. The result was that AI
344. 11:56training only ran at 11% of full
345. 11:58capacity according to sources. Seeing
346. 12:00that it was useless for training, XAI
347. 12:03decided to rent out the data center and
348. 12:05move the training of Grok to a new one.
349. 12:07So, that's not a great sign in
350. 12:08management competence. But okay, that's
351. 12:10fine. The company can still make money
352. 12:12with a data center in space.
353. 12:17SpaceX has filed with the FCC to build a
354. 12:19space cloud of up to 1 million
355. 12:21satellites. Orbital AI compute is
356. 12:24targeted to begin by 2028, but launching
357. 12:26a million satellites is a lot. The
358. 12:28Starlink network has 9,000 satellites.
359. 12:31So, imagine over 100 times that number
360. 12:33orbiting the Earth.
361. 12:35>> [music]
362. 12:36>> Two years ago, I literally moderated a
363. 12:38panel on space junk, and I was even on
364. 12:40the news talking about it.
365. 12:42Although SpaceX's satellites have
366. 12:44autonomous avoidance systems,
367. 12:45coordinating 9,000 objects and
368. 12:48coordinating 1 million objects are two
369. 12:50completely different challenges. I truly
370. 12:52believe that if we don't want space
371. 12:54debris becoming a hot topic in 15 to 20
372. 12:56years, sending 1 million satellites into
373. 12:58orbit with GPU technology that will be
374. 13:01outdated in 3 years just isn't a good
375. 13:04idea. Or am I the crazy one here? Maybe
376. 13:07there's something I'm missing, but it
377. 13:08doesn't look like a financially
378. 13:10competitive idea at all. In fact, it
379. 13:12seems kind of wasteful.
380. 13:14It seems like even in the best case
381. 13:15scenario, with every new GPU release
382. 13:17from Nvidia, the orbital data center
383. 13:19will become less and less competitive
384. 13:21over time. In the worst case, if LLM
385. 13:24algorithms get vastly more efficient or
386. 13:26AI running locally on devices becomes
387. 13:28the norm, that seems to me like a huge
388. 13:30threat to the central thesis of a space
389. 13:32data center.
390. 13:33But I do get the basic idea here. On my
391. 13:36last episode about data centers, I got
392. 13:38so many comments from you guys saying
393. 13:39how horrible it is living next to them.
394. 13:41Some of your stories were heartbreaking.
395. 13:43So yes, I can see people wanting an
396. 13:45alternative. But as for actually putting
397. 13:47data centers in space, I'm just not so
398. 13:48sure personally. So it remains to be
399. 13:50seen how well this works out
400. 13:52economically.
401. 13:54From an engineering point of view,
402. 13:55leaving aside the cost of launching that
403. 13:57much mass, dissipating the heat
404. 13:59radiation from the GPUs will be a
405. 14:00challenge. In space, there's no air, no
406. 14:03water, and no convection. So radiating
407. 14:06heat away becomes very inefficient. But
408. 14:08this is their area of expertise, so
409. 14:10they've probably got it figured out.
410. 14:12That being said, there are some other
411. 14:14things to consider though. They are as
412. 14:16follows.
413. 14:17Cosmic rays and protons can cause bit
414. 14:19flipping in memory, resulting in errors
415. 14:21in calculations, and that's not great
416. 14:23for some of the most densely packed
417. 14:25transistor technology in existence. And
418. 14:27also, energetic solar particles can
419. 14:29destroy unshielded GPUs. Van Allen belt
420. 14:32electrons degrade solar panels over
421. 14:34time. And most devastating of all, solar
422. 14:36flares can cause massive radiation
423. 14:37sparks, resulting in the potential loss
424. 14:39of the GPUs. To fix this, each satellite
425. 14:43must require heavy radiation shielding
426. 14:45and possibly redundant systems. Without
427. 14:47it, the data center performance will
428. 14:49inevitably degrade over time. All of
429. 14:50this add mass and cost.
430. 14:53But what do I know? Obviously, I'm not
431. 14:55sure what the future holds. So we'll
432. 14:56have to wait and see if this risky long
433. 14:58shot plays out. So, if orbital data
434. 15:01centers don't end up working, in the
435. 15:02prospectus, they're still talk of
436. 15:04asteroid mining and Mars colonies. So
437. 15:06that's got to be worth a lot of money,
438. 15:07right?
439. 15:08Well, even so, SpaceX only believes that
440. 15:11it's a subset of 15% of their market.
441. 15:13Remember, 85% is AI. So when you think
442. 15:17about it, the company themselves don't
443. 15:18really believe that. It's a bit of a
444. 15:20shame that AI is now 85% of SpaceX's
445. 15:23addressable market. This is essentially
446. 15:25what investors are buying according to
447. 15:27their own filings.
448. 15:28>> [music]
449. 15:28>> So, what will happen to the stock?
450. 15:30Well, nobody really knows, but the
451. 15:32general consensus is that it's going to
452. 15:33rocket up as it launches. Then there'll
453. 15:35be a slow grind downwards.
454. 15:39SpaceX
455. 15:39>> [music]
456. 15:39>> was a genuine, world-class aerospace
457. 15:42company with a profitable satellite
458. 15:43internet business that had no
459. 15:45competition. If these were normal days,
460. 15:47they would have IPO'd on the basis of
461. 15:48steady growth. They would have zero
462. 15:50competition, so they could have jacked
463. 15:52up their prices as much as they wanted.
464. 15:54That should have been enough for
465. 15:55investors. But everyone caught the AI
466. 15:57bug, and now, for some reason, AI has to
467. 16:00be 85% of their business.
468. 16:02xAI is essentially a Trojan horse
469. 16:05because in the prospectus, you have to
470. 16:06scroll through 11 pages of rockets and
471. 16:08talk of reaching the stars before you
472. 16:10get to the financial details that reveal
473. 16:12that the whole plan [music] is to float
474. 16:13an AI company. Investors seem none the
475. 16:16wiser.
476. 16:17>> I heard last time they checked, it was
477. 16:19like 200 something. They're probably
478. 16:21going to go over to $500.
479. 16:22>> 10 million?
480. 16:23>> Hundreds of thousands of dollars.
481. 16:25>> 35?
482. 16:26>> Really? What is he selling? He's selling
483. 16:28real estate on [music] Mars.
484. 16:30>> The S&P 500 thankfully said no to this,
485. 16:33but the Nasdaq said yes. So, look, who
486. 16:35knows? I could be completely wrong. A
487. 16:37rabbit could be pulled out of a hat, and
488. 16:39SpaceX could rise to a hundred trillion
489. 16:41dollar valuation in five years. But as
490. 16:43for right now, it seems very [music]
491. 16:45risky, and there's also a bad financial
492. 16:47smell. Anyway, I'm done talking about
493. 16:49it. I'm just going to get the popcorn.
494. 16:51Right. So, thanks for watching. What do
495. 16:53you guys think about the SpaceX IPO? Do
496. 16:55you think things are going to go great
497. 16:56and it's going to make a lot of money,
498. 16:57or do you think it's just a quick,
499. 16:59[music] panicked cash grab off the back
500. 17:00of the AI boom? Let me know your
501. 17:02thoughts in the comment section below.
502. 17:04Feel free to subscribe if you liked it.
503. 17:05[music] My name is Dagogo. You've been
504. 17:07watching ColdFusion, and I'll catch you
505. 17:09again soon for another episode. Cheers,
506. 17:10[music] guys. Have a good one.
507. 17:18>> [music]
508. 17:23[music]
509. 17:31[music]
510. 17:41>> Cold fusion.
511. 17:43It's me thinking.