WEBVTT 00:00:00.340 --> 00:00:01.460 Good afternoon, everyone. 00:00:02.000 --> 00:00:04.820 Thank you for joining us today for SpaceX's second quarter 00:00:04.820 --> 00:00:06.700 2026 earnings conference call. 00:00:07.340 --> 00:00:09.160 I'm Andrea Williams, head of Investor Relations. 00:00:09.940 --> 00:00:12.280 Joining me today are Elon, Gwen, and Brett, 00:00:12.860 --> 00:00:15.500 and we are speaking to you from our facility in Bastrop, Texas. 00:00:16.490 --> 00:00:18.290 Our second quarter financial results were announced 00:00:18.290 --> 00:00:20.190 just after 3 p.m. Central Time, 00:00:20.590 --> 00:00:22.470 and all relevant materials have been published 00:00:22.470 --> 00:00:26.170 to our Investor Relations website at ir.spacex.com. 00:00:26.760 --> 00:00:29.280 For this webcast, our prepared remarks will be followed 00:00:29.280 --> 00:00:30.840 by a question and answer session. 00:00:31.560 --> 00:00:33.340 We will take live Q&A from equity analysts, 00:00:34.140 --> 00:00:35.980 and then move to questions directly from our investors 00:00:36.340 --> 00:00:39.100 that were polled on our dedicated earnings Q&A platform. 00:00:40.380 --> 00:00:42.540 Thank you to everyone who submitted your questions in advance. 00:00:43.650 --> 00:00:45.810 During this webcast, we will discuss our business outlook 00:00:45.810 --> 00:00:47.370 and make forward-looking statements. 00:00:48.030 --> 00:00:49.310 These statements are based on our predictions 00:00:49.310 --> 00:00:50.870 and expectations as of today. 00:00:51.820 --> 00:00:54.400 Actual events or results could differ materially 00:00:54.400 --> 00:00:55.960 due to a number of risks and uncertainties, 00:00:56.520 --> 00:00:58.800 including those mentioned in our most recent filings 00:00:58.800 --> 00:01:01.700 with the SEC, including our final prospectus 00:01:01.700 --> 00:01:03.660 and today's published Form 10Q. 00:01:04.600 --> 00:01:05.860 Additionally, during this call, 00:01:06.360 --> 00:01:07.960 we may discuss certain non-GAP measures. 00:01:08.800 --> 00:01:10.440 In our press release in our filings with the SEC, 00:01:10.740 --> 00:01:12.180 you can find additional disclosures, 00:01:12.680 --> 00:01:15.660 including all reconciliations with comparable GAP measures. 00:01:16.860 --> 00:01:18.520 For the second quarter of 2026, 00:01:19.160 --> 00:01:20.960 we announced the following financial results 00:01:20.960 --> 00:01:23.000 as compared to the prior year quarter. 00:01:23.960 --> 00:01:28.580 Revenues of $7.8 billion, up 92%, 00:01:28.580 --> 00:01:34.240 from $4.1 billion, a quarterly net loss of $541 million, 00:01:35.260 --> 00:01:38.100 an improvement of $467 million, 00:01:38.920 --> 00:01:41.620 an adjusted EBITDA of $3.5 billion, 00:01:42.860 --> 00:01:46.680 up 191% from $1.2 billion. 00:01:47.850 --> 00:01:49.090 Now I will hand the call over to Elon 00:01:49.090 --> 00:01:50.130 for his opening remarks. 00:01:50.710 --> 00:01:55.190 Thank you. 00:02:00.460 --> 00:02:02.680 So it's been another milestone year for SpaceX so far. 00:02:03.260 --> 00:02:06.240 We're making great progress developing 00:02:06.840 --> 00:02:10.300 a fully and rapidly reusable reliable rocket with Starship. 00:02:11.360 --> 00:02:13.380 We're connecting millions more people globally 00:02:13.380 --> 00:02:17.400 across 170 markets with fast reliable styling internet. 00:02:18.300 --> 00:02:21.100 We're building AI compute capacity at scale, 00:02:21.360 --> 00:02:22.940 fast than anyone else would believe. 00:02:23.900 --> 00:02:25.980 And we're significantly improving our AI models, 00:02:26.940 --> 00:02:30.260 most notably with the release of Grog 4.5 last month. 00:02:30.700 --> 00:02:33.880 And now I'll dive into each of those topics separately. 00:02:35.340 --> 00:02:38.120 So with Starship, we completed two successful flights 00:02:38.380 --> 00:02:41.500 of Starship B3 in the past 90 days. 00:02:42.100 --> 00:02:44.560 Flight 13 demonstrated core capabilities necessary 00:02:44.560 --> 00:02:46.020 to achieve an orbital mission 00:02:46.300 --> 00:02:47.860 and return to Starbase for catch. 00:02:48.940 --> 00:02:51.160 And on Flight 14 will be our first flight 00:02:51.560 --> 00:02:56.180 to fly our version three styling satellites, 00:02:56.340 --> 00:02:58.680 our communication satellites to operational orbit. 00:03:01.190 --> 00:03:05.030 We will try to actually catch the first and the second stage 00:03:05.030 --> 00:03:07.370 of the Starlink B3 this year. 00:03:08.250 --> 00:03:11.050 And we could possibly catch the ship as soon as the next flight. 00:03:14.900 --> 00:03:20.210 And it's difficult to actually explain to most people 00:03:20.210 --> 00:03:22.910 just the incredible significance of Starship. 00:03:25.320 --> 00:03:28.400 And I think perhaps describing it in terms of tonged orbit 00:03:28.400 --> 00:03:32.140 is a good way to appreciate the magnitude 00:03:32.880 --> 00:03:38.230 of the impact Starship will have on the future of civilization. 00:03:39.530 --> 00:03:45.590 So right now we deliver roughly 2,500 tons a year 00:03:46.030 --> 00:03:48.310 to orbit via Falcon. 00:03:49.850 --> 00:03:52.810 And this is a big number by normal standards. 00:03:53.430 --> 00:03:56.530 The rest of the world I think live is around 300 tons. 00:03:57.990 --> 00:04:03.300 So we deliver 80 to 90% of total Earth 00:04:03.460 --> 00:04:06.140 mass to orbit per year with Falcon. 00:04:06.720 --> 00:04:11.430 Now with Starship, our aspirations, 00:04:11.530 --> 00:04:12.950 and I think we will achieve these aspirations 00:04:13.730 --> 00:04:19.470 are to deliver well over a million tons to orbit per year 00:04:20.110 --> 00:04:23.970 and probably ultimately 10 million tons per year. 00:04:25.530 --> 00:04:28.170 So this is not like if you had a chart 00:04:28.970 --> 00:04:32.950 where you plotted things you would need a very, 00:04:32.950 --> 00:04:37.170 very big table to put the chart on to even see 00:04:38.710 --> 00:04:40.650 any of our competitors on that chart 00:04:40.650 --> 00:04:43.370 and to make that sound next biggest competitor 00:04:43.370 --> 00:04:45.710 was one pixel high on the bar chart. 00:04:46.230 --> 00:04:48.310 You would still need a very big bar chart 00:04:48.310 --> 00:04:52.270 to represent the SpaceX Starship vehicle. 00:04:55.270 --> 00:04:58.130 And it's worth noting that we actually encourage 00:04:58.130 --> 00:04:59.610 our competitors to copy us. 00:04:59.810 --> 00:05:01.810 We do nothing to slow them down. 00:05:01.810 --> 00:05:05.670 We actually launch, we launch competing satellite constellations 00:05:06.260 --> 00:05:09.350 at the fair prices, the same price we charge everyone. 00:05:10.170 --> 00:05:13.130 So we do encourage our competitors to copy us 00:05:13.130 --> 00:05:15.910 but it's quite difficult to copy something like Starship. 00:05:17.630 --> 00:05:24.330 But we don't use patents as a sort of a land, 00:05:24.530 --> 00:05:26.390 mine exercise to stop people from copying us. 00:05:26.510 --> 00:05:27.410 We actively encourage it. 00:05:28.030 --> 00:05:31.850 So, anyways, but when you think about it on that, 00:05:32.030 --> 00:05:35.240 so that if we go from 2,500 tons a year to 00:05:36.300 --> 00:05:38.820 a million potentially 10 million tons per year, 00:05:39.000 --> 00:05:42.340 it's really a ridiculously profound difference. 00:05:43.540 --> 00:05:47.540 But that is our plan and I think we will achieve that plan. 00:05:49.500 --> 00:05:54.730 Regarding Starlink, it's also worth just noting 00:05:54.730 --> 00:05:59.450 that the Starlink V3 satellite is about an order of magnitude 00:05:59.450 --> 00:06:02.410 more capable than the Starlink V2 satellite, 00:06:02.750 --> 00:06:04.990 which was already a very capable satellite. 00:06:05.990 --> 00:06:09.770 And we expect to launch about an order of magnitude 00:06:10.210 --> 00:06:11.830 more Starlink V3 satellites. 00:06:13.390 --> 00:06:17.550 So that would mean a roughly two order of magnitude 00:06:18.130 --> 00:06:19.630 increase in the delivered bandwidth. 00:06:21.510 --> 00:06:26.890 Even if our monetization per bit dropped by a factor of 10, 00:06:27.210 --> 00:06:30.590 that would still mean a 10x increase in the revenue of Starlink. 00:06:31.420 --> 00:06:34.600 So I think people are really underestimating Starlink here. 00:06:38.030 --> 00:06:41.360 So this is a big deal. 00:06:41.640 --> 00:06:43.400 It's kind of hard for people to wrap their minds around this, 00:06:43.620 --> 00:06:48.320 but it's not out of the question that at some point 00:06:49.320 --> 00:06:52.360 Starlink will deliver a majority of the world's internet, 00:06:53.700 --> 00:06:56.120 at least in countries where we're allowed to operate, 00:06:56.300 --> 00:06:57.880 which is the vast majority of countries. 00:07:00.140 --> 00:07:02.900 So this is an important bear in mind, 00:07:02.960 --> 00:07:04.500 and it's not in the infinity future, 00:07:05.160 --> 00:07:09.370 it's less than 10 years. 00:07:14.190 --> 00:07:18.470 And with the advent of AI and humanoid robotics 00:07:18.470 --> 00:07:21.970 and vehicle robotics and just a massive number of robots, 00:07:22.390 --> 00:07:24.690 the appetite for bandwidth will be much greater 00:07:24.690 --> 00:07:25.890 than it has been in the past. 00:07:26.430 --> 00:07:28.850 You think of a human consuming, perhaps, 00:07:29.230 --> 00:07:33.850 or producing perhaps a few hundred bits per second of output, 00:07:34.650 --> 00:07:39.310 and actually the average human is outputting less than one bit 00:07:39.310 --> 00:07:42.890 per second over the course of a day, which is 86,400 seconds. 00:07:44.490 --> 00:07:49.860 Now computers easily do billions of bits per second 00:07:49.860 --> 00:07:53.500 continuously over a 24-hour period. 00:07:53.640 --> 00:07:58.120 So this is why I would expect the appetite for bandwidth 00:07:58.120 --> 00:08:01.160 to grow dramatically with the growth of AI and robotics. 00:08:03.400 --> 00:08:05.020 And really, I think Starlink is the only thing 00:08:05.020 --> 00:08:07.240 that can actually service that bandwidth. 00:08:10.900 --> 00:08:12.000 Regarding AI, 00:08:13.910 --> 00:08:17.290 we're making rapid progress on Grog, 00:08:17.990 --> 00:08:20.550 because Grog 4.5 was a huge improvement, 00:08:21.230 --> 00:08:25.650 and we have Grog 4.6 coming out probably next week, 00:08:26.470 --> 00:08:32.150 and then Grog 4.7 is about three or four weeks from today, 00:08:32.450 --> 00:08:36.670 and we expect the cadence of AI development 00:08:36.670 --> 00:08:41.250 to improve dramatically, and then with Grog 5, 00:08:41.250 --> 00:08:44.490 which should be out before the end of this year, 00:08:45.070 --> 00:08:48.610 we will be incorporating the entire corpus of SpaceX data, 00:08:48.830 --> 00:08:52.570 basically all the data that SpaceX has ever produced, 00:08:53.590 --> 00:08:56.050 which is a tremendous amount of the course of a quarter century, 00:08:56.530 --> 00:08:58.210 will be incorporated into Grog training. 00:08:59.110 --> 00:09:01.390 And we think this will make Grog by far the best engineer. 00:09:03.710 --> 00:09:07.650 So these are all like very, very big things. 00:09:08.310 --> 00:09:11.590 In addition, of course, we are providing compute to others, 00:09:12.400 --> 00:09:16.000 and we are building and deploying compute, 00:09:16.740 --> 00:09:20.320 I think faster, our rate of growth certainly is faster 00:09:20.320 --> 00:09:21.280 than anyone else. 00:09:21.740 --> 00:09:24.980 And our efficiency of compute deployment I think is also the highest. 00:09:27.650 --> 00:09:32.330 So we expect to end this year with over two gigawatts of compute, 00:09:33.090 --> 00:09:38.470 and probably our cumulative compute online 00:09:39.320 --> 00:09:43.320 by the end of next year will be several times higher. 00:09:45.180 --> 00:09:51.720 So it may, let's say, be closer to 10 gigawatts of compute 00:09:51.720 --> 00:09:53.080 than five gigawatts of compute. 00:09:54.980 --> 00:09:55.700 There's a way to think about it. 00:09:56.940 --> 00:10:00.600 And going forward, we've decided to build exclusively on Nvidia 00:10:01.540 --> 00:10:06.870 because we think the Veraruban architecture 00:10:06.870 --> 00:10:08.230 is the best architecture. 00:10:09.070 --> 00:10:10.550 We think it's the best AI computer. 00:10:11.860 --> 00:10:15.600 And we greatly value our close cooperation 00:10:15.600 --> 00:10:18.900 and partnership on many levels with Nvidia. 00:10:19.580 --> 00:10:22.420 So we're exclusive to Nvidia. 00:10:25.720 --> 00:10:31.180 And this new capacity coming online will obviously further benefit 00:10:31.180 --> 00:10:34.440 the development of GARC as well as our cursor acquisition, 00:10:35.320 --> 00:10:38.400 which should close quite soon. 00:10:39.220 --> 00:10:41.640 I think we're through almost all of the regulatory hurdles. 00:10:44.500 --> 00:10:49.260 And then with respect to the star-mind AI satellite, 00:10:49.820 --> 00:10:54.240 which will be essentially an optimized Veraruban 00:10:54.240 --> 00:10:56.980 NVL 72 computer. 00:10:58.700 --> 00:11:05.300 This is not some sort of far future distant thing we expect 00:11:05.300 --> 00:11:07.760 to start launching these next year. 00:11:10.430 --> 00:11:23.960 And we think the design of the NVL 72 VR computer is a much better 00:11:23.960 --> 00:11:29.800 design than having a standard RAC-style design. 00:11:30.800 --> 00:11:34.340 And so we expect to actually deploy this on the ground as well as in orbit 00:11:35.660 --> 00:11:40.540 because we think it's going to be a radical simplification 00:11:40.540 --> 00:11:44.620 of the normal NVL 72 RAC. 00:11:45.320 --> 00:11:48.480 It will cost less, it will be more effective. 00:11:50.460 --> 00:11:51.800 And so if we're going to put it in space, 00:11:52.020 --> 00:11:52.720 we want to put it on the ground. 00:11:54.480 --> 00:11:56.340 So I think that's going to be pretty cool. 00:11:59.870 --> 00:12:03.210 In conclusion, the SpaceX team is solving some of the hardest 00:12:03.210 --> 00:12:04.850 engineering problems in the history of humanity. 00:12:06.740 --> 00:12:11.240 And I think the team is succeeding because we've got, 00:12:13.320 --> 00:12:19.240 in a lot of ways, the smartest, most dedicated team of humans 00:12:19.240 --> 00:12:20.860 that has ever existed. 00:12:21.840 --> 00:12:24.580 I'm incredibly proud to work with such a team 00:12:24.580 --> 00:12:28.920 and I'd like to thank the team for their incredible hard work. 00:12:30.320 --> 00:12:33.740 And also thank you to investors who have supported us long way. 00:12:39.040 --> 00:12:39.820 Thanks, Elon. 00:12:40.830 --> 00:12:43.070 We had an exceptional second quarter, 00:12:43.070 --> 00:12:46.550 and today I'll focus on the operational and commercial highlights 00:12:46.550 --> 00:12:50.010 as well as the initiatives we are prioritized going forward. 00:12:50.970 --> 00:12:52.570 Starting with our space business, 00:12:52.910 --> 00:12:55.030 we remain the leading launch provider globally 00:12:55.730 --> 00:13:01.470 with 78 total launches and 1,041 tons of mass to orbit delivered 00:13:01.470 --> 00:13:02.910 in the first half of this year, 00:13:03.610 --> 00:13:06.930 primarily allocated to our own internal Starlink missions, 00:13:07.470 --> 00:13:09.790 but with great customer missions as well. 00:13:10.470 --> 00:13:14.050 While our launch capabilities are key to supporting our connectivity business, 00:13:14.370 --> 00:13:17.290 we continue to have a very robust commercial manifest 00:13:18.030 --> 00:13:21.330 and a very robust and growing US government mission demand, 00:13:21.590 --> 00:13:23.130 which drives revenue for the segment. 00:13:23.770 --> 00:13:26.970 We're currently launching at our highest Falcon cadence 00:13:26.970 --> 00:13:29.910 and are on the precipice of operationalizing Starship. 00:13:31.070 --> 00:13:34.370 As Elon said, the first two flights of our V3 Starship vehicles 00:13:34.370 --> 00:13:37.810 put us in an excellent position to achieve our near term goals 00:13:37.810 --> 00:13:41.570 of reaching orbit, catching and reusing the ships and boosters 00:13:41.570 --> 00:13:44.630 and deploying our V3 Starlink broadband satellites. 00:13:46.540 --> 00:13:51.060 Speaking of which, Starlink had a standout quarter driven by continued growth 00:13:51.060 --> 00:13:55.060 across our consumer enterprise and government businesses. 00:13:55.760 --> 00:14:00.200 During the second quarter, we added net more than 1.7 million 00:14:00.200 --> 00:14:02.920 Starlink subscribers globally, consumer. 00:14:03.600 --> 00:14:06.340 This reflects our best quarter of new customers to date 00:14:06.340 --> 00:14:09.820 and increase from the 1.4 million we added during the first quarter. 00:14:10.950 --> 00:14:13.270 We achieved this by holding ARPU, 00:14:13.710 --> 00:14:16.070 which is the average revenue per user, 00:14:16.730 --> 00:14:19.910 stable at $66 per month compared to the first quarter. 00:14:19.910 --> 00:14:24.890 We ended the second quarter with service availability in 167 markets 00:14:24.890 --> 00:14:28.390 and have continued to activate new markets since then 00:14:28.390 --> 00:14:31.290 with some very exciting announcements coming soon, 00:14:31.290 --> 00:14:32.950 maybe even as early as today. 00:14:34.530 --> 00:14:37.830 As of June 30th, our constellation of operational Starlink broadband 00:14:37.830 --> 00:14:41.770 and mobile satellites in orbit grew to roughly 10,200, 00:14:42.490 --> 00:14:47.830 with our 9,600 broadband satellites delivering roughly 800 terabits per second 00:14:47.830 --> 00:14:49.710 of total downlink capacity. 00:14:50.650 --> 00:14:56.130 Following a successful speed run test of the V3 Starlink satellites 00:14:56.130 --> 00:14:58.170 on the most recent Starship launch, 00:14:58.670 --> 00:15:02.310 during which we connected every satellite through every laser link, 00:15:02.890 --> 00:15:06.510 we intended to deploy V3 satellites into the constellation 00:15:06.510 --> 00:15:10.110 for operational use on the upcoming Starship missions. 00:15:11.050 --> 00:15:13.650 The significant amount of capacity we're able to add 00:15:13.650 --> 00:15:16.050 to the Starlink constellation from the V3 satellites 00:15:16.050 --> 00:15:19.630 will enable us to continue providing even better service. 00:15:20.210 --> 00:15:21.370 I think it's pretty great already, 00:15:21.910 --> 00:15:24.930 but to do so while serving more and more customers over the world. 00:15:25.450 --> 00:15:26.890 In fact, in the years ahead, 00:15:27.090 --> 00:15:30.510 we expect Starlink will represent a significant portion of global internet traffic, 00:15:30.790 --> 00:15:31.950 which Elon also talked about. 00:15:32.950 --> 00:15:35.910 We also see significant growth opportunities for Starlink 00:15:35.910 --> 00:15:38.950 on the enterprise and government side of the business. 00:15:39.610 --> 00:15:41.870 We believe revenue from these markets have the potential 00:15:41.870 --> 00:15:46.990 to reach a scale at least comparable but likely exceed our consumer business 00:15:46.990 --> 00:15:48.650 as we continue to gain share. 00:15:49.330 --> 00:15:52.430 In the second quarter, we signed a major agreement with American Airlines, 00:15:52.950 --> 00:15:56.190 as well as activated our service with new partners, including Southwest, 00:15:57.170 --> 00:15:59.250 Virgin Atlantic, Iberia, and Aerlingus. 00:15:59.870 --> 00:16:02.670 We have become the provider of choice among major airlines 00:16:02.670 --> 00:16:07.130 with outstanding customer feedback, but still have significant room for growth. 00:16:07.130 --> 00:16:11.150 In fact, we heard from one of our airline customers, 00:16:11.430 --> 00:16:19.470 shockingly customers are flying shorter hop flights instead of direct 00:16:20.030 --> 00:16:23.450 so that they could ensure that they're on a Starlink activated flight. 00:16:23.930 --> 00:16:25.630 They'd never seen this in the business before. 00:16:27.380 --> 00:16:32.420 On the government side, we won more than $6 billion in U.S. contracts in Q2, 00:16:32.700 --> 00:16:37.480 supporting major space force programs that offer our nation mission critical communications 00:16:37.480 --> 00:16:41.960 and sensing capabilities, and we see even more room for growth in this sector 00:16:42.480 --> 00:16:43.620 in this coming year. 00:16:44.940 --> 00:16:47.580 In Q2, we also launched the new Starlink mobile partnerships 00:16:47.580 --> 00:16:53.560 with international carriers, including SoftBank, NTT, Dokomo, and Spark New Zealand. 00:16:54.120 --> 00:16:57.460 Our near-term priority is launching mobile V2 satellites on Starship 00:16:57.460 --> 00:17:03.100 ahead of integrating the 65 megahertz of Echostar spectrum later next year. 00:17:03.920 --> 00:17:06.840 This spectrum transfer was recently approved by the FCC 00:17:06.840 --> 00:17:11.100 and represents a foundational competitive advantage for Starlink mobile. 00:17:12.300 --> 00:17:16.320 In AI, the industry is hurtling towards ever more compute demands, 00:17:16.920 --> 00:17:21.240 we're rapidly expanding our compute capacity to meet our own needs, as GRC expands, 00:17:21.660 --> 00:17:25.000 as well as those for other leading companies such as Google and Anthropic. 00:17:26.060 --> 00:17:30.360 In terms of our own GRC models, the enterprise feedback on rollout of GRC 4.5 00:17:30.360 --> 00:17:34.780 has been very positive so far, token consumption tripled out of the gate 00:17:34.780 --> 00:17:38.880 after the July release, and we continue to see strong signals on usage 00:17:38.880 --> 00:17:40.740 and monetization for the model. 00:17:41.500 --> 00:17:43.860 We're looking forward to welcoming the cursor team to SpaceX 00:17:43.860 --> 00:17:48.500 to integrate our engineering and begin to benefit from a combined sales capability. 00:17:49.500 --> 00:17:52.540 In review, our momentum as a business is extraordinary. 00:17:53.180 --> 00:17:56.180 Our engineering teams are scaling Starship and our next-gen satellites, 00:17:56.180 --> 00:18:00.680 while our sales teams expand Starlink, aviation, mobility, and global markets. 00:18:01.400 --> 00:18:03.900 At the same time, we're bringing on additional compute capacity 00:18:03.900 --> 00:18:09.080 and expanding and deepening our AI partnerships across a broader set of customers. 00:18:10.060 --> 00:18:12.920 And though I've been in this business for 24 years, 00:18:13.040 --> 00:18:15.940 I could not be more excited about the months ahead 00:18:15.940 --> 00:18:19.020 with all the exciting and amazing things going on. 00:18:19.280 --> 00:18:22.320 It really feels like we're just getting started all over again. 00:18:25.470 --> 00:18:26.910 Now I'll turn the call over to Brett. 00:18:26.910 --> 00:18:28.700 Thank you, Gwen. 00:18:29.810 --> 00:18:32.070 We demonstrated strong results in Q2. 00:18:32.770 --> 00:18:35.470 Revenue growth accelerated across every segment. 00:18:36.170 --> 00:18:38.370 We narrowed our net losses by nearly half, 00:18:38.790 --> 00:18:42.450 and on a year-over-year basis, on a year-over-year basis, 00:18:42.710 --> 00:18:45.630 and we grew adjusted even meaningfully faster than revenue. 00:18:46.570 --> 00:18:49.410 Importantly, we significantly strengthened the balance sheet 00:18:49.410 --> 00:18:52.830 through our IPO and inaugural investment grade bond offering. 00:18:52.830 --> 00:18:56.510 The combination of our global launch leadership, 00:18:57.010 --> 00:19:00.950 recurring Starlink revenue and rapid growth in AI infrastructure 00:19:00.950 --> 00:19:04.530 continues to drive improved operating leverage. 00:19:05.250 --> 00:19:07.470 Now I'll walk through the financials for each segment. 00:19:08.250 --> 00:19:12.410 For the space segment, revenue grew 55% sequentially 00:19:12.410 --> 00:19:16.490 and 29% year-over-year to 962 million, 00:19:17.030 --> 00:19:20.390 driven by a higher number of larger customer launches 00:19:20.390 --> 00:19:22.990 and a favorable customer mix. 00:19:23.690 --> 00:19:27.670 Segment costs and expenses rose 389 million year-over-year, 00:19:28.050 --> 00:19:30.470 and we continued to accelerate R&D investments 00:19:30.470 --> 00:19:31.690 in our Starship program. 00:19:32.770 --> 00:19:36.030 Starship aims to quadruple payload capacity 00:19:36.030 --> 00:19:38.650 and reduce launch costs by 10 times 00:19:38.650 --> 00:19:40.290 compared to our Falcon 9 rocket, 00:19:40.770 --> 00:19:44.810 unlocking significant capabilities across all of our business segments. 00:19:45.660 --> 00:19:48.340 We continue to make progress building the infrastructure 00:19:48.340 --> 00:19:52.400 required to support thousands of Starship launches per year, 00:19:53.180 --> 00:19:56.320 including accelerating Raptor and launch vehicle production, 00:19:57.100 --> 00:19:58.760 the build-out of our Gigabays, 00:19:59.240 --> 00:20:03.340 and making significant progress towards activating multiple launch pads 00:20:03.340 --> 00:20:05.920 at Starbase and in Cape Canaveral, 00:20:06.540 --> 00:20:08.760 at Pad 39A and Pad 37. 00:20:09.970 --> 00:20:13.750 Space-adjusted EBITDA was a loss of 205 million for the quarter, 00:20:14.170 --> 00:20:16.650 primarily reflecting these higher R&D expenses. 00:20:16.650 --> 00:20:20.510 Moving to the connectivity segment, 00:20:21.150 --> 00:20:23.910 revenue was up in the connectivity segment by, 00:20:24.430 --> 00:20:29.570 or was up to 4.3 billion, up 32% sequentially, 00:20:29.810 --> 00:20:31.390 and 66% year-over-year. 00:20:32.140 --> 00:20:35.420 Growth was driven by record net additions of Starlink subscribers 00:20:35.420 --> 00:20:39.040 and a sharp increase in enterprise and government revenue. 00:20:40.040 --> 00:20:44.800 Starlink subscriber ARPU was $66 unchanged compared to Q1. 00:20:44.800 --> 00:20:47.700 As we continue our global expansion, 00:20:48.360 --> 00:20:50.740 we're taking a localized go-to-market approach 00:20:50.740 --> 00:20:54.160 that ensures our product and service offerings fit local needs. 00:20:55.110 --> 00:20:59.250 While our geographic expansion may drive down blended ARPU over time, 00:20:59.690 --> 00:21:04.270 we anticipate continued subscriber momentum to deliver strong revenue growth. 00:21:05.460 --> 00:21:09.220 Enterprise and government revenue grew 108% year-over-year 00:21:09.220 --> 00:21:11.980 and represents a durable source of revenue, 00:21:11.980 --> 00:21:15.040 contributing to strong segment margins overall. 00:21:16.010 --> 00:21:19.650 As Gwen mentioned, we see enormous opportunity for growth in this segment, 00:21:20.130 --> 00:21:21.630 as we remain, for example, 00:21:22.290 --> 00:21:25.630 less than 10% penetrated in one of the largest segments being aviation. 00:21:28.800 --> 00:21:31.340 Total costs and expenses for the connectivity segment 00:21:31.340 --> 00:21:35.860 increased 970 million, or approximately 58% year-over-year. 00:21:36.600 --> 00:21:40.640 The increase was driven by higher spend to support our revenue growth, 00:21:40.640 --> 00:21:43.940 including growth in our satellite constellation, 00:21:44.720 --> 00:21:48.860 increased R&D investment for our next-gen version 3 satellites, 00:21:49.500 --> 00:21:51.060 and incremental marketing spend. 00:21:52.230 --> 00:21:56.510 As Elon mentioned, our version 3 satellites are expected to deliver 10 times 00:21:56.510 --> 00:22:00.730 the amount of broadband capacity and data density from our current generations, 00:22:01.430 --> 00:22:03.470 which I think we believe are already pretty amazing, 00:22:03.930 --> 00:22:07.230 and which are essential to supporting continued subscriber growth 00:22:07.230 --> 00:22:09.230 and higher value service tiers. 00:22:10.690 --> 00:22:15.360 Income from operations from the connectivity segment increased 79% year-over-year 00:22:15.360 --> 00:22:19.500 to $1.7 billion, outpacing revenue growth 00:22:19.500 --> 00:22:22.940 and driving nearly three points of operating margin expansion. 00:22:24.460 --> 00:22:30.140 Connectivity segment adjusted EBITDA rose 64% year-over-year to $2.6 billion. 00:22:32.440 --> 00:22:38.060 AI segment revenue was $2.6 billion, up 213% sequentially, 00:22:38.060 --> 00:22:40.540 and 247% year-over-year. 00:22:41.390 --> 00:22:44.810 The increase was driven primarily by new cloud services agreements, 00:22:45.630 --> 00:22:49.110 as well as growth in GROC and X subscription revenue. 00:22:49.770 --> 00:22:55.710 Advertising revenue grew 7% sequentially, as we overhauled our advertising platform technology 00:22:55.710 --> 00:23:01.130 to enable AI functionality and enhance self-service for all customer types. 00:23:03.450 --> 00:23:06.250 In Q2, we entered into cloud services agreements 00:23:06.250 --> 00:23:11.770 providing access to compute capacity at our Colossus and Colossus 2 sites. 00:23:12.570 --> 00:23:17.230 The initial ramp from these agreements contributed $1.6 billion of incremental 00:23:17.230 --> 00:23:19.950 AI infrastructure revenue in the quarter. 00:23:21.210 --> 00:23:26.190 Total cost and expenses for the AI segment rose $1.6 billion year-over-year, 00:23:26.770 --> 00:23:31.030 reflecting increased R&D primarily due to higher infrastructure spend 00:23:31.030 --> 00:23:33.310 as we accelerated compute deployment. 00:23:33.310 --> 00:23:39.030 We meaningfully narrowed our AI segment net operating loss to $1.3 billion 00:23:39.030 --> 00:23:45.690 and turned adjusted EBITDA positive for the segment in the second quarter at $1.1 billion. 00:23:46.950 --> 00:23:52.330 The incremental revenue from new hosting deals generated high incremental EBITDA margins 00:23:52.330 --> 00:23:54.870 as we monetized available compute capacity. 00:23:55.620 --> 00:24:02.340 We ended the second quarter with 1.4 gigawatts of nameplate compute up from 1 gigawatt in Q1 00:24:02.340 --> 00:24:05.320 and 400 megawatts a year earlier. 00:24:06.180 --> 00:24:10.820 We expect to end this year at over 2 gigawatts of compute capacity. 00:24:12.490 --> 00:24:17.810 Total company capital expenditures in the second quarter were approximately $18.4 billion, 00:24:18.470 --> 00:24:22.730 of which roughly $15.8 billion supported AI compute infrastructure. 00:24:23.430 --> 00:24:28.070 The remainder funded ongoing Starship and launch infrastructure, satellite production, 00:24:29.110 --> 00:24:30.870 and global ground station expansion. 00:24:30.870 --> 00:24:35.040 Additionally, we paid out $856 million during the quarter 00:24:35.040 --> 00:24:39.480 under our spectrum credit agreement related to our pending EcoStar transaction. 00:24:41.410 --> 00:24:44.590 Our capital allocation priorities remain clear and disciplined. 00:24:45.390 --> 00:24:49.390 We plan to continue to invest aggressively in the three areas we believe 00:24:49.390 --> 00:24:51.630 generate the highest long-term returns. 00:24:52.350 --> 00:24:56.110 Starship development and production scale, next generation, 00:24:56.110 --> 00:25:01.010 star link, broadband and mobile constellations, and AI compute infrastructure. 00:25:02.120 --> 00:25:05.800 Across all three of these investments, we remain focused on capital efficiency, 00:25:06.560 --> 00:25:11.140 return on investment, and preservation of a strong investment grade balance sheet. 00:25:12.860 --> 00:25:16.220 The second quarter was transformative from a capital markets perspective. 00:25:16.840 --> 00:25:22.400 We completed our initial public offering, raising approximately $85.7 billion in net proceeds, 00:25:22.400 --> 00:25:29.000 and followed that with a $25 billion inaugural investment grade senior notes offering 00:25:29.000 --> 00:25:32.420 that was partially used to repay our $20 billion bridge loan. 00:25:33.700 --> 00:25:37.880 The senior notes were issued across five tranches with a weighted average interest rate 00:25:37.880 --> 00:25:42.240 of 5.855% and average maturity of 11.7 years. 00:25:43.250 --> 00:25:47.450 We ended the quarter with $100 billion of cash, cash equivalents, and marketable securities, 00:25:48.250 --> 00:25:50.390 and $47.5 billion in backlog. 00:25:50.390 --> 00:25:57.670 Looking ahead, we continue to see robust demand in all three of our business segments, 00:25:58.230 --> 00:26:00.530 but in particular in our cloud services arrangements. 00:26:01.490 --> 00:26:05.190 We see increasingly favorable economics with each agreement we sign, 00:26:05.690 --> 00:26:11.330 and as Elon mentioned, we expect the supply, demand, and balance in the compute market to continue. 00:26:12.040 --> 00:26:15.880 The current economics have translated into a less than one year payback 00:26:15.880 --> 00:26:17.960 on our new capital deployments for compute. 00:26:18.890 --> 00:26:21.490 For example, in the first few weeks of the third quarter, 00:26:21.970 --> 00:26:26.910 we've already contracted an additional $6.7 billion of cloud services revenue 00:26:26.910 --> 00:26:31.450 over a six month period that begins ramping starting in October of this year. 00:26:32.320 --> 00:26:37.420 We believe this puts us on a trajectory, including contribution from cursor, 00:26:38.020 --> 00:26:44.140 to reach 100 billion of ARR or annualized revenue run rate by the end of this year, 00:26:44.140 --> 00:26:48.160 based on our expected revenue in the month of December of this year. 00:26:49.340 --> 00:26:51.880 With that, I'll turn the call back to Andrea. 00:26:54.430 --> 00:26:56.090 Thank you, Elon, Gwen, and Brett. 00:26:57.030 --> 00:26:58.390 This concludes our prepared remarks. 00:26:59.230 --> 00:27:02.930 For today's Q&A session, we will take our first questions from the equity analysts. 00:27:03.650 --> 00:27:08.490 Please press star one to join the question queue, and also please limit yourself to one question. 00:27:09.110 --> 00:27:11.010 Operator, please proceed with the queue. 00:27:13.050 --> 00:27:15.710 We will now begin the question and answer session. 00:27:15.710 --> 00:27:20.370 If you would like to ask a question, please press star one to raise your hand. 00:27:21.010 --> 00:27:23.490 To withdraw your question, press star one again. 00:27:24.370 --> 00:27:28.850 We ask that you pick up your handset when asking a question to allow for optimum sound quality. 00:27:29.770 --> 00:27:32.890 If you are muted locally, please remember to unmute your device. 00:27:33.850 --> 00:27:36.170 Please stand by while we compile the Q&A roster. 00:27:41.400 --> 00:27:45.120 Your first question comes from the line of Eric Sheridan with Goldman Sachs. 00:27:45.700 --> 00:27:47.280 Your line is open. Please go ahead. 00:27:48.440 --> 00:27:52.720 Thanks so much for taking the question and thanks for all the details in the prepared remarks. 00:27:53.280 --> 00:27:54.940 Maybe just one on the connectivity segment. 00:27:55.260 --> 00:28:00.380 Can you discuss the backlog of enterprise and government contracts in that segment of the business 00:28:00.380 --> 00:28:05.380 and how we think about that backlog converting it to revenue and what that might mean for the piece 00:28:05.380 --> 00:28:07.720 of cadence of growth in the next couple of years? 00:28:08.060 --> 00:28:08.460 Thanks so much. 00:28:12.410 --> 00:28:16.770 Yeah, we are quite bullish on our enterprise activities. 00:28:17.450 --> 00:28:25.410 I mentioned alone that we received from the government over $6 billion in contracts. 00:28:26.910 --> 00:28:33.810 You can look at those as tranches and beginnings of additional capability that we will bring online. 00:28:34.770 --> 00:28:41.730 Some of which we will have to compete, but fundamentally we have the contracts in place to fulfill many 00:28:41.730 --> 00:28:50.350 phenomenologies, both the ones we recently announced as well as additional ones that we will be working on going forward. 00:28:51.370 --> 00:28:52.950 Enterprise revenue is quite sticky. 00:28:53.230 --> 00:28:55.090 We have never lost an enterprise customer. 00:28:55.390 --> 00:28:59.730 They are quite happy with the Starlink capability that we have right now. 00:29:00.290 --> 00:29:04.270 Brett mentioned we are only 10% penetrated in the aviation industry, 00:29:05.070 --> 00:29:09.550 which definitely represents a lot of headroom for revenue going forward. 00:29:09.550 --> 00:29:15.810 Maritime, I think, is Starlink opens up the TAM dramatically. 00:29:16.890 --> 00:29:25.510 Starlink offers the capability that ships that don't really have much connectivity right now because they can't afford VSATs. 00:29:25.910 --> 00:29:29.330 We can basically outfit those ships with Starlink maritime. 00:29:29.810 --> 00:29:38.750 I am quite bullish on enterprise, very bullish on government and quite happy with the stickiness that we have seen in those markets. 00:29:40.200 --> 00:29:49.380 I will just elaborate a little bit on that, which is that obviously in order to secure a large amount of enterprise revenue, 00:29:49.700 --> 00:29:53.420 you have to demonstrate a very high reliability, very high uptime. 00:29:55.160 --> 00:29:57.680 In the early days, Starlink did not have high uptime. 00:29:58.580 --> 00:30:01.020 There are a lot of enterprise customers, most enterprise customers, I would say, 00:30:01.080 --> 00:30:06.280 that still have that early Starlink experience where it was a little patchy. 00:30:07.160 --> 00:30:17.900 We are building up a large enterprise sales team just to go out there, meet with enterprise corporate and government customers, 00:30:18.760 --> 00:30:25.440 show them the incredible uptime and low latency of the Starlink system today, 00:30:26.100 --> 00:30:33.340 so they realize that they can actually treat Starlink as a primary provider, not sort of as a backup provider. 00:30:34.780 --> 00:30:42.760 This requires just going out there, talking to people, educating them about the currency of the Starlink system and where it's going, 00:30:44.140 --> 00:30:48.360 and I would expect enterprise revenue to substantially exceed consumer revenue. 00:30:53.430 --> 00:30:54.470 Next question, operator. 00:30:55.050 --> 00:30:58.230 Your next question comes from the line of Adam Jonas with Morgan Stanley. 00:30:58.850 --> 00:31:00.230 Your line is open. Please go ahead. 00:31:01.860 --> 00:31:03.600 Thanks. Elon, first one for you. 00:31:03.600 --> 00:31:11.880 Just confirming, you said you expect to be closer to 10 gigawatts of nameplate compute by end of 27, closer to 10 than 5. 00:31:12.880 --> 00:31:14.360 Do you have line of sight? 00:31:14.920 --> 00:31:19.200 How common is your line of sight on permitting chips and turbines for that? 00:31:19.640 --> 00:31:21.160 And then I have a follow up for Gwen. Thanks. 00:31:24.640 --> 00:31:33.400 So we're actually aiming to far exceed that gigawatt number in terms of power online, power cooling, 00:31:33.400 --> 00:31:37.320 and electrical equipment. 00:31:38.180 --> 00:31:46.540 So our tentative target is to actually have 20 gigawatts at the power and cooling level online by the end of next year. 00:31:46.780 --> 00:31:49.360 Now, I don't think we're going to achieve 20 gigawatts per week. 00:31:49.520 --> 00:32:03.380 We want to have a series of projects that cumulatively come to 20 gigawatts by the end of next year, 00:32:04.380 --> 00:32:06.360 something close to 15 gigawatts. 00:32:07.980 --> 00:32:13.760 So assuming that maybe a quarter of the projects take longer than expected. 00:32:14.580 --> 00:32:21.220 So our goal is to have far more power, cooling, and electrical equipment than we have GPUs. 00:32:22.140 --> 00:32:26.600 That's the logical thing to do, given the relative expense of GPUs versus balance of system. 00:32:28.920 --> 00:32:31.360 SpaceX's company is incredibly good at hardware. 00:32:32.460 --> 00:32:38.300 In fact, I'd say probably outside of China, SpaceX and Tesla are the two best companies on Earth at hardware. 00:32:41.260 --> 00:32:43.660 Yeah, and I think perhaps even including China. 00:32:43.980 --> 00:32:53.840 So we're taking a small amount of the expertise that we use for rockets and satellites and applying that to scaling terrestrial data centers. 00:32:54.820 --> 00:33:04.420 And we're finding even a small amount of what we've learned building rockets, which are incredibly difficult, applied to data centers, yields tremendous amounts. 00:33:06.900 --> 00:33:14.010 So yeah, we expect to have far an excess of the power and cooling that's needed. 00:33:14.010 --> 00:33:25.590 And then our understanding within the video is that we will receive a very significant percentage of their GPUs next year. 00:33:28.680 --> 00:33:31.340 Okay, thanks, Elon. Just Glenn, just a follow up. 00:33:32.460 --> 00:33:40.500 You were recently asked what is one of the more under appreciated parts of the SpaceX investment story and you said robotics. 00:33:41.800 --> 00:33:43.480 Can you elaborate on why you believe that? 00:33:43.480 --> 00:33:47.640 How do robots fit into the SpaceX TAM? Thanks, Glenn. 00:33:50.340 --> 00:33:53.260 Let me try to put something together for you. 00:33:54.240 --> 00:34:00.300 I think what I might have said is under appreciating the amount of connectivity that we will need going forward. 00:34:01.090 --> 00:34:14.700 And I think robotics, especially robotics tied to AI, like for autos and humanoid robots, will dramatically increase the demand for connectivity dramatically. 00:34:15.460 --> 00:34:17.560 Elon talked about that in his remarks. 00:34:18.870 --> 00:34:20.890 I mean, you can think about satellites as being robots. 00:34:21.250 --> 00:34:23.110 I mean, they're autonomous and they do not need servicing. 00:34:26.070 --> 00:34:32.760 So, and then, you know, in the sort of what may seem the far future, but we'll come faster than you think. 00:34:33.900 --> 00:34:42.420 You know, using robots on the moon to scale up manufacturing on the moon, which it sounds like super sci-fi right now, but it's going to happen. 00:34:43.700 --> 00:34:48.580 We'll enable us to, well, the mass accelerator on the moon. 00:34:49.260 --> 00:35:09.020 And if you have a mass accelerator on the moon and you have solar and radiator production on the moon, you can, I know this sounds totally nuts, but you can probably scale to a thousand times the economy of Earth in terms of intelligence launched to space. 00:35:09.880 --> 00:35:13.000 But probably, you know, maybe even a million times. 00:35:15.430 --> 00:35:18.760 So, you know, we are going to land a lot of time on the moon. 00:35:19.800 --> 00:35:21.280 We're going to build the factories on the moon. 00:35:21.660 --> 00:35:22.800 The robots will be helpful with that. 00:35:23.660 --> 00:35:31.940 And then it's like, well, how much intelligence would you like as we, you know, grow it towards being a Kardashev 2 scale civilization. 00:35:34.020 --> 00:35:35.380 Did that get that for you, Adam? 00:35:41.370 --> 00:35:44.530 Your next question comes from the line of John Godin with Citigroup. 00:35:44.910 --> 00:35:45.650 Your line is open. 00:35:45.930 --> 00:35:46.490 Please go ahead. 00:35:48.990 --> 00:35:50.970 Hey, thanks for taking my question. 00:35:51.410 --> 00:36:03.910 One of the hallmarks of SpaceX today to tell efficiently the organization has used Capitol so far in its journey, but it's also no secret that the Capitol needs their significant to support your vision from here. 00:36:04.310 --> 00:36:05.990 Obviously, the company's moving fast. 00:36:05.990 --> 00:36:15.090 So, I was hoping you could talk a bit more about the CAPEX outlook across the segment, touch upon any areas where buying birth building might accelerate growth. 00:36:15.430 --> 00:36:17.650 The mobile strategy obviously comes to mind there. 00:36:18.310 --> 00:36:26.410 And any related thoughts on Capitol needs for next steps in extreme vertical integration like Terrapab, anything that's needed to support future growth. Thanks. 00:36:28.550 --> 00:36:30.510 Yeah, I appreciate the question. 00:36:30.830 --> 00:36:36.670 I think we are very proud of how we have managed capital allocation over the years for sure. 00:36:38.110 --> 00:36:41.810 And I think we are continuing down that path, right? 00:36:41.990 --> 00:36:48.090 If you look at it, we have been very efficient to date and I think will continue to be. 00:36:48.530 --> 00:36:59.990 If you look at the rest of this year to answer specifically on CAPEX, I think you should probably think that the next two quarters are very similar to the current quarter from a CAPEX level perspective. 00:36:59.990 --> 00:37:08.910 But I do think as you look at longer term, I think CAPEX really, I think for our perspective is all CAPEX is not the same, right? 00:37:09.010 --> 00:37:14.830 What we're seeing right now is a situation where we're able to deploy specifically on the AI compute side. 00:37:15.070 --> 00:37:19.410 We're able to deploy capital in such a way that we're getting less than a one year payback. 00:37:19.850 --> 00:37:27.970 It's almost like your capital is becoming almost an item that you would put otherwise in cogs because it's moving so quickly. 00:37:27.970 --> 00:37:30.370 And as far as monetization and moving into revenue. 00:37:30.890 --> 00:37:39.490 That's different than I would say historically in the SpaceX level where we're building assets that are going to benefit us from launch sites for the next couple of decades. 00:37:40.330 --> 00:37:47.990 And so we certainly manage for optimizing ROI on our capital all the time from that perspective. 00:37:49.720 --> 00:37:51.680 Yeah, I'll add a little bit to that. 00:37:53.060 --> 00:38:00.780 As Brett mentioned, we are expecting to reach 100 billion plus ARR in December of this year. 00:38:02.300 --> 00:38:12.860 It's probably also worth mentioning that our internal projections for reaching a trillion dollars in revenue, not ARR, but revenue have moved up from 2031 to 2030. 00:38:13.680 --> 00:38:22.340 So prior to the IPO, you know, the financial projections we had were reaching a trillion dollars in revenue in 2031. 00:38:22.620 --> 00:38:24.400 We now expect that to be in 2030. 00:38:26.300 --> 00:38:29.140 And there's a nonzero chance of that being in 2029. 00:38:31.470 --> 00:38:35.570 Your next question comes from the line of Ronald Epstein with Bank of America. 00:38:36.190 --> 00:38:37.730 Your line is open. Please go ahead. 00:38:39.200 --> 00:38:41.540 Yeah, hey, good evening, everybody. 00:38:42.440 --> 00:38:48.440 Could you walk through what you learned on Starship launch 13? 00:38:49.180 --> 00:38:51.840 You know, what was better than expected, you know, what was worse? 00:38:52.580 --> 00:38:58.680 And then, you know, how do we think about this in the journey to kind of rapid reusability, you know, the heat shield, so on, so forth? 00:39:01.600 --> 00:39:02.000 Sure. 00:39:02.260 --> 00:39:09.560 So the, yeah, actually Flight 13, I wasn't worried about that Flight 13, you know, because, you know, think of Apollo 13, like things went wrong on that flight. 00:39:09.560 --> 00:39:12.960 But this flight actually went incredibly well. 00:39:13.880 --> 00:39:16.640 The heat shield appears to be very robust. 00:39:17.480 --> 00:39:20.100 The ship is still floating in the ocean. 00:39:21.600 --> 00:39:26.120 And we expect to recover the ship and bring it back for analysis. 00:39:27.960 --> 00:39:30.280 So I'd say things look very good. 00:39:32.420 --> 00:39:38.640 And that's why we, assuming we receive regulatory approval to do so, 00:39:38.640 --> 00:39:45.840 we'll attempt to catch the ship with the tower on the next flight, which is tentatively scheduled for the end of this month. 00:39:46.740 --> 00:39:50.540 We expect the cadence flights to be increasing rapidly. 00:39:54.030 --> 00:40:02.570 And probably a year from now, we will be doing at least one flight a day, possibly more. 00:40:05.160 --> 00:40:09.020 Your next question comes from a line of Doug Anmouth with JP Morgan. 00:40:09.440 --> 00:40:11.040 Your line is open. Please go ahead. 00:40:12.840 --> 00:40:14.440 Thanks so much for taking questions. 00:40:15.500 --> 00:40:31.220 Elon, can you elaborate on the key unlocks from leveraging your rockets and satellite expertise just in accelerating the pace of compute build out, and then totally appreciate the $6.7 billion in incremental compute deals. 00:40:31.760 --> 00:40:44.120 Can you talk more about just how you expect the supply demand environment to persist going forward and whether you'll be able to keep this kind of premium pricing for these deals and sustainability? 00:40:44.980 --> 00:40:45.180 Thanks. 00:40:48.500 --> 00:40:48.900 Sure. 00:40:49.380 --> 00:40:54.220 Well, you know, as the saying goes, you know, long as they say like they would say like this ain't rocket science. 00:40:54.520 --> 00:40:56.500 Well, actually rocket science is literally our daily business. 00:40:58.520 --> 00:41:05.340 And, you know, rocket science is a idiomatic expression for extreme technological difficulty. 00:41:05.600 --> 00:41:11.840 And there's a reason for it because the thing, let me tell you what rockets desperately want to do every flight. 00:41:12.100 --> 00:41:14.100 They desperately want to blow themselves into tiny pieces. 00:41:15.500 --> 00:41:22.140 And then the engineering struggle is to convince the rocket not to blow itself into tiny pieces and actually deliver payload orbit. 00:41:22.140 --> 00:41:26.480 But you'll note that, you know, some of our competitors do, in fact, blow their rockets up. 00:41:27.200 --> 00:41:31.360 This is not a reflection on their does not mean they're in cable. 00:41:31.360 --> 00:41:36.020 In fact, I think our competitors are very smart people, but despite being very smart, the rockets will blow up. 00:41:37.640 --> 00:41:51.490 So now if you apply a small amount of the engineering skill that goes into making gigantic rockets reach orbit on a regular basis reliably, 00:41:52.570 --> 00:42:00.370 to threshold data centers, this is like, frankly, the Yankees, New York Yankees going in and playing a little league team. 00:42:00.850 --> 00:42:03.690 And it's kind of ridiculous, frankly. 00:42:05.550 --> 00:42:14.150 The threshold data centers are a trivial problem compared to making gigantic reusable rockets and which are launched frequently. 00:42:15.530 --> 00:42:20.370 And the Starlink satellites also are like alien level technology. 00:42:22.210 --> 00:42:31.290 There's still nothing that compares to our Starlink V2 satellites and our V3 satellites are really more than an order of magnitude better than our V2 satellites. 00:42:32.710 --> 00:42:38.590 So you take a small percentage of that engineering applied to threshold data centers and you get an amazing outcome. 00:42:40.290 --> 00:42:41.930 So that's what we're seeing. 00:42:46.070 --> 00:43:01.380 Now, from a monetization standpoint, if you say like, what's my best guess for monetization per watt of, say, Rubens, my guess is 00:43:02.140 --> 00:43:04.360 would be somewhere between $30 and $50. 00:43:06.780 --> 00:43:10.300 This is just a guess. I'm just, you know, take a guess as a guess. 00:43:12.910 --> 00:43:18.750 But what we're seeing with AI is that the intelligence per watt is increasing rapidly. 00:43:20.050 --> 00:43:23.150 So that means the usefulness per watt is increasing rapidly. 00:43:24.390 --> 00:43:34.830 And I think anyone who's a close poller of the AI arena, you're starting smaller and smaller models do more and more amazing things. 00:43:36.270 --> 00:43:50.450 There's an argument that we may get to the point where even an H100 is able to deliver useful AI that's comparable to a human engineer. 00:43:52.620 --> 00:43:56.480 Based on current trends, I think that's not an impossible outcome. 00:43:57.720 --> 00:43:59.380 So then it's like how useful would that be? 00:43:59.780 --> 00:44:03.120 I think the usefulness that would be is going to be extremely high. 00:44:03.120 --> 00:44:10.030 I mean, if you think of like, where was AI 12 months ago? 00:44:11.360 --> 00:44:12.400 I mean, this is 12 months ago. 00:44:12.560 --> 00:44:20.800 It would have been before the advent of cloud 4.5, which I think is should be considered one of the milestones and the credit to entrophic. 00:44:22.480 --> 00:44:27.560 You know, it was, I would say a shock to the system was how good cloud 4.5 was. 00:44:30.890 --> 00:44:33.590 And that was, you know, 10 months ago, basically. 00:44:36.050 --> 00:44:39.450 And if you say, well, what about an AI two years ago? 00:44:39.610 --> 00:44:42.130 I think, well, an AI two years ago feels like something that should be in a museum. 00:44:42.610 --> 00:44:42.790 Okay. 00:44:43.170 --> 00:44:45.370 Well, you're like, you're not going to believe this, but two years ago, this was AI. 00:44:47.470 --> 00:44:47.990 This relic. 00:44:51.170 --> 00:45:00.040 So I think, you know, by the end of next year, it's not clear to me that there's anything that's happening. 00:45:00.040 --> 00:45:06.940 Anything that digital at least that AI won't be able to do based on the current rate of improvement. 00:45:09.270 --> 00:45:15.400 Then look at the rate at which logic and memory is being produced. 00:45:17.280 --> 00:45:19.340 And one was always considered a limiting factor here. 00:45:19.420 --> 00:45:20.700 A limiting factor currently is memory. 00:45:21.770 --> 00:45:27.810 The memory output is increasing by around 20% per year. 00:45:28.650 --> 00:45:38.820 Now, normally that would be fantastically fast and amazing for any large mature industry. 00:45:39.920 --> 00:45:43.520 But ask yourself, is the demand increasing by 20% year? 00:45:43.680 --> 00:45:45.580 No, the demand is increasing by 200% a year. 00:45:46.100 --> 00:45:46.400 Maybe higher. 00:45:48.610 --> 00:45:56.910 So if you've got demand increasing much faster than supply, then economics 101 would suggest that 00:45:56.910 --> 00:45:58.790 the price increases. 00:46:00.400 --> 00:46:01.120 It does not decrease. 00:46:04.720 --> 00:46:08.340 Your next question comes from the line of Edison U with Deutsche Bank. 00:46:08.680 --> 00:46:09.440 Your line is open. 00:46:09.740 --> 00:46:10.340 Please go ahead. 00:46:12.510 --> 00:46:14.610 Thank you for taking our question. 00:46:14.990 --> 00:46:19.750 Wanted to come back on the comment about the 100 billion in AR. 00:46:20.350 --> 00:46:27.470 Is there any way you can can unpack that a little bit and perhaps kind of what it takes to get there? 00:46:28.230 --> 00:46:34.010 And then in relation to that, Elon, I know you mentioned you pulled the revenue target forward by year. 00:46:34.610 --> 00:46:39.370 Can you maybe call out maybe a few things that gave you this extra confidence? 00:46:39.890 --> 00:46:40.130 Thank you. 00:46:43.780 --> 00:46:50.420 Well, on the unpacking, I think that the key there is it's really gross in all three of the business segments. 00:46:51.120 --> 00:46:55.100 But the biggest piece is, and we already talked about, you know, additional business. 00:46:55.100 --> 00:47:01.360 Since we've already closed in the first couple of weeks of July on the AI cloud services side. 00:47:01.760 --> 00:47:12.180 But certainly, as you know, we've already announced some previous deals with Google and entropic that start to ramp either later this quarter or, you know, in October. 00:47:12.920 --> 00:47:23.700 And so there's a lot of momentum from that side of the business as well as we're seeing, you know, strength in our own models and and cursor coming into the fold as well. 00:47:23.700 --> 00:47:27.720 So it's really all three segments that are driving that. 00:47:27.860 --> 00:47:32.260 But the cloud services piece being the largest contributor of that piece. 00:47:34.080 --> 00:47:38.740 Yes, like to be clear, the 100 billion ARR in December is not a question mark. 00:47:39.080 --> 00:47:41.500 That's what we'd achieve if we basically did nothing. 00:47:42.780 --> 00:47:47.510 So like, you know, so I think it may be higher than that. 00:47:48.030 --> 00:47:49.410 It probably will be higher than that. 00:47:49.410 --> 00:48:02.940 The, you know, I think in terms of the relative percentage of, you know, breakdown of like what revenue generates what. 00:48:04.260 --> 00:48:08.660 I think the breakdown is probably similar to what we shared pre IPO. 00:48:09.340 --> 00:48:12.560 It's just happening a year sooner. 00:48:15.650 --> 00:48:29.440 And, you know, I will go back to what I said and what Gwen and Brett mentioned earlier, which is that we do expect a very rapid increase in revenue from strong and communications from the V3 satellite. 00:48:30.250 --> 00:48:44.020 Because like, like I said, it's more than 10 times the capability of the V2 satellite and we'll be launching 10 times as many and we have the direct to sell consolation going up as well. 00:48:44.900 --> 00:49:01.460 So, you know, so some of you are looking at the sort of historical growth of Starlink based on V2, but we have a step change, a massive step change happening with the V3 satellite. 00:49:01.460 --> 00:49:11.130 Your next question comes from the line of John Hodelick with UBS. Your line is open. Please go ahead. 00:49:12.640 --> 00:49:17.680 Great. Thank you. Two things. First, you want to just follow up on that comment on Starlink mobile. 00:49:18.540 --> 00:49:33.780 When do you expect to start launching the V2s? He talked about your appetite for additional spectrum after the upper star deal and then just how big do you think that the opportunity is in mobile and if you could compare that to the broadband business, that would be great. 00:49:33.780 --> 00:49:39.880 And then secondly, just a quick follow up on the compute. Looks like you have a big ramp and compute coming in 2027. 00:49:40.780 --> 00:49:54.480 How should we think of that capacity being leased versus used for, you know, to train Grog? I guess in other words, how much of that has to be withheld to train Grog versus that can be leased out to third parties? Thanks. 00:49:56.660 --> 00:50:07.560 I'll go ahead and take the Starlink mobile questions, John. We're going to start flying the next generation Starlink mobile satellite, which is confusing. 00:50:07.880 --> 00:50:16.980 It's actually the second version as opposed to the V3, which is the broadband, next generation broadband satellite. We're going to start flying those next year. 00:50:16.980 --> 00:50:25.220 And just so you understand currently the system that we're operating uses about five megahertz of bandwidth. 00:50:26.680 --> 00:50:42.800 And we operate that through local telco providers. We have 65 megahertz of bandwidth available to us through the EcoStar spectrum, which gives massive increase in capability. 00:50:43.340 --> 00:50:59.520 Another way to look at it is we were also probably 10x the number of satellites. So we, simplifying, you could look at the Starlink mobile leveraging the next generation satellite and the EcoStar spectrum is being 100 times better, 00:51:00.160 --> 00:51:08.140 10 times 10 better than what we've got right now. And it's a pretty decent service. You can do voice and video calls over signal and WhatsApp. 00:51:08.820 --> 00:51:18.760 So we anticipate really strong uptake of that capability. As far as revenue, like I'm just going to take a very high level kind of top down. 00:51:19.300 --> 00:51:36.240 The big three in the United States, AT&T, Verizon, and TMO roughly between them, $600 billion a year. 00:51:36.240 --> 00:51:53.720 Better. We will eliminate dead zones leveraging the, basically, the satellites in orbit. It will be better during any sort of natural disaster because, surprisingly, even those space movies make space looks super dangerous is a pretty quiet environment. 00:51:55.840 --> 00:52:04.860 And so I'm quite excited about Starlink mobile going forward. We will start to fly the satellites next year, and we will start providing service end of next year. 00:52:06.880 --> 00:52:20.440 Yeah, I guess I should elaborate on, we obviously have more than one question being answered, but the, you know, obviously, if somebody's trying to create a financial model, some of these questions are, you have to plug in some number here. 00:52:21.120 --> 00:52:33.220 So for Grock training, we do expect the percentage of compute that we use internally for rock training to decrease over time and to have far more of the compute used for inference. 00:52:34.000 --> 00:52:38.360 Or to, you know, for, for ranting training inference compute to others. 00:52:39.640 --> 00:52:46.970 So I would expect maybe, you know, to a 10% of our compute to be used for Grock training, something like that. 00:52:50.000 --> 00:52:56.520 Your next question comes from the line of CATGUN moral with Evercore ISI. Your line is open. Please go ahead. 00:52:58.860 --> 00:53:05.580 Good afternoon, and thanks for taking the question. I had a follow up on Starlink mobile and your ambitions in the US market in particular. 00:53:05.580 --> 00:53:22.820 I think a lot of us are trying to better understand what a US direct-to-consumer service could look like to become a true fourth US carrier would imply well over 100 billion of mobile CAPX over time, much of it, auction spectrum, and green field infrastructure. 00:53:23.640 --> 00:53:36.680 How are you weighing that potential bill plan against the MBO know or acquisition path? And what's your internal base case on how far you'll take this envision, or is this all a little bit too premature at this point? Thank you. 00:53:39.840 --> 00:53:51.340 The spectrum that we purchase from Echo Star does have terrestrial components. So we definitely intend to build out terrestrial, the terrestrial component. So you will have not only the, 00:53:51.340 --> 00:54:08.500 the capacity from the satellites themselves, but you will have a build out of the terrestrial, basically the hardware and systems necessary to make a true mobile service exactly what you want it to be. 00:54:09.300 --> 00:54:18.000 I'm not going to talk about the CAPX for that. It's a little, well, I don't want to say it's immature. I'm just going to say I'm not going to talk about the CAPX for that right now. 00:54:18.800 --> 00:54:32.840 But let's talk a little, let me give a little hint, and that is you could put a base station, a cellular base station on basically the gear that holds a Starlink broadband dish. 00:54:33.420 --> 00:54:47.320 So you can have kind of these little femto cells around the country, and you deploy that as you need it. You don't have to necessarily spend, you know, many billions of dollars in low band spectrum 00:54:47.320 --> 00:55:01.260 all up front before you deploy your system. So that's why I don't want to talk about the CAPX for building out the terrestrial because we have a lot of really kind of great and new ideas for how we're going to do that. And I think it will be quite CAPX efficient. 00:55:03.400 --> 00:55:30.270 Yeah, that point is probably worth just a few more words, which is that because this is quite an important point. Instead of having to deploy these very expensive and difficult to locate large cellular base stations, we feel reasonably confident that we can deploy a large number of sort of small 00:55:30.270 --> 00:55:39.370 small stations, essentially, that are really just Starlink dishes that also provide connectivity in the mobile spectrum bands and have them be all over the place. 00:55:41.150 --> 00:55:57.690 And the Starlink antennas are located on the roofs of houses and businesses. So they're, you know, it's sort of a clear view for providing connectivity directly to the cell phones on the ground. 00:55:57.690 --> 00:56:06.400 And, in fact, providing connectivity that is, I think, probably better and higher bandwidth than what is currently available from cellular providers. 00:56:07.450 --> 00:56:09.890 This is an important point. 00:56:13.700 --> 00:56:19.380 Thank you. With the remaining time, we would like to cover some of the questions from individual shareholders. I will read them aloud. 00:56:21.300 --> 00:56:26.260 We got a lot of questions on the progress for Starship human landing system. Can we provide any updates? 00:56:30.740 --> 00:56:50.180 For Starship, well, obviously we need to make Starship super reliable for satellites before we put any people on board, but we are expecting to have a very high launch rate unless achieve a level of safety needed for humans very quickly, probably by the end of next year. 00:56:50.180 --> 00:57:06.360 As far as the milestones go, propellant transfer in orbit is critical to both our internal SpaceX as well as our HLS ambitions. 00:57:07.420 --> 00:57:19.690 Artemis 3 mission is next year. That's where we will dock with the Orion spacecraft. We will follow that up with a direct to lunar cargo mission on crude. 00:57:19.690 --> 00:57:26.650 And then we want to put boots on the ground, boots on the moon in 2028. 00:57:30.740 --> 00:57:37.720 For the cursor acquisition, there were several questions asking about the combined product roadmap for the companies once it closes. 00:57:43.520 --> 00:57:46.020 I think we will leave that to... 00:57:48.280 --> 00:57:52.660 We will probably make some announcement or discussion of that outside of the owning school. 00:57:54.140 --> 00:57:59.040 But we don't want to jump the gun on regulators for the closing acquisition. 00:57:59.960 --> 00:58:04.040 So we are trying to get the acquisition close as quickly as possible. I think we are pretty close to that. 00:58:04.760 --> 00:58:08.300 But we have to worry of sort of jumping the gun on regulatory closures. 00:58:10.800 --> 00:58:18.730 Next question from the individual shareholders. Once you start launching broadband V3 satellites, how quickly will they translate into improved Starlink service? 00:58:22.240 --> 00:58:29.770 So would we need some critical mass of V3 satellites, probably at least on the order of about 1,000? 00:58:32.830 --> 00:58:44.120 But probably that's approximately second quarter next year. I think we would get to that point. 00:58:48.100 --> 00:58:52.920 And our last question, is the heat shield the hardest remaining problem for Starship? 00:58:55.760 --> 00:59:04.440 Look, I don't want to jinx it or anything, but I think I would consider the heat shield problem solved at this point. 00:59:06.100 --> 00:59:13.080 We want to take a close look at the ship that is currently floating in the ocean. 00:59:14.560 --> 00:59:26.820 But all indications from data and our visual inspection suggest that we have solved the heat shield problem, 00:59:28.560 --> 00:59:34.960 which is arguably the single biggest problem. That doesn't mean we want to make improvements. 00:59:35.280 --> 00:59:38.820 The heat shield, of course, will continue to make improvements to the heat shield. 00:59:39.000 --> 00:59:46.020 But I would say that we do not see any technical obstacles at this point to achieving full unwrapped reasability. 00:59:49.610 --> 00:59:56.030 Thank you. This concludes today's call. You can reach us at ir.spacex.com, and we will see you next quarter. 00:59:56.350 --> 00:59:56.570 Cheers. 00:59:57.880 --> 01:00:01.840 This concludes today's call. Thank you for attending. You may now disconnect.