00:00The idea that SpaceX is undertaking TerraFab with Tesla at unprecedented scale to secure compute here on Earth and in
00:09orbit, your reaction to that, Karen?
00:13Well, my reaction is they've actually been thinking about this for a long time.
00:18So Lauren just talked about the FCC application for a million sats in space.
00:24But if you rewind back to 2022, they had filed a patent around getting computing resources to sats in space
00:32as well.
00:32So I think this is something that Elon and his team have been thinking about for a long time, frankly,
00:38at least since the post-chat GPT moment.
00:41I also think conceptually it makes a lot of sense.
00:45So you just talked about, Lauren just talked about harnessing power in sun-synchronous orbit.
00:51But the other constraints on AI data centers here on Earth are not just power, it's land, it's permitting, it's
01:00getting contractors like electricians to put these boxes together.
01:04And those constraints don't exist in space.
01:07And so I think conceptually it makes a ton of sense to be thinking about in the longer term data
01:14centers in space.
01:15The question, of course, becomes what you were just talking about.
01:18Is it economical?
01:19And frankly, what we've been watching for as investors is Starship.
01:26And Starship Flight 12 should happen in April.
01:29They're going to be launching their Gen 3 of the ship, which will carry up to 100 tons of payload.
01:37And as a reminder, the plan is for Starship to be fully reusable in the future.
01:42And if it's fully reusable, ultimately the cost of launching it could come down very, very meaningfully, which would mean
01:51launching the AI data centers into space, those payloads, could make that whole concept very economical.
01:57Karen, you and your team at Fidelity, what research and modeling have you done for the factors and the point
02:03at which orbital data center becomes profitable?
02:06A lot of people, when I said you were coming on the program, wanted to know if you've been shown
02:11any evidence by SpaceX about the viability of an orbital data center this decade.
02:17But many people measure it in the dollar per kilogram equation when it comes to Starship.
02:23Absolutely.
02:24We do measure it in dollar per kilogram.
02:26And again, as I mentioned, it really comes down to how quickly Starship can be up and fully reusable.
02:35So as a reminder, Falcon 9, they just reused the booster, the first stage.
02:40And Starship will be both the booster and the ship.
02:44They've now successfully caught the booster of Starship three times at the tower catch.
02:48People have probably seen that, and they've had successful atmospheric reentry of the ship as well, but have not yet
02:55caught that.
02:56The other thing that they haven't done as of yet is they actually haven't dispensed of any actual payloads into
03:02space as of yet.
03:03They've tested out the PES door, which opens up to let the payloads out, but actual payloads have not been
03:10put out into space.
03:12So those are the milestones that we're really following.
03:15But you can bet there is an Excel spreadsheet at Fidelity that measures the cost per kilowatt.
03:22And when we believe this will become economical.
03:24Oh, there's going to be people who really want to see that Excel spreadsheet, Karen.
03:28In particular, those who look more cynical on the outlook of the economy and the economic pros and cons of
03:35this.
03:35For example, Jim Chanos, well known for perhaps betting against certain key themes, but he asks in a post in
03:43particular how much one gigawatt data center in space would save in terms of power costs versus one gigawatt terrestrial
03:49data center.
03:49Can you push out where you think it could go at least?
03:53Yeah, I mean, listen, I think, again, we're talking not this year, not next year.
03:57We are talking about several years out, but it could move down the cost pretty meaningfully.
04:04I won't give you an exact dollar amount, but we do believe it could be pretty material.
04:10In the here and now, what we're all looking for materially is up to a $1.25 trillion valuation if
04:15SpaceX does indeed go public.
04:18How realistic do you think that is at the moment?
04:19What momentum do we need to see?
04:21Because at the moment, we're all sort of at the behest of the markets.
04:24We are at the behest of the market.
04:27Obviously, the markets are feeling quite good this morning as well.
04:31But I think it's important to remember that SpaceX has a lot going for it right now.
04:38You know, when we first invested at Fidelity back in 2015 in SpaceX, they had done 13 Falcon 9 launches
04:45to date.
04:46Today, they launch every other day.
04:49They're putting up somewhere between 80% to 90% of the world's payload.
04:54One company, when we first invested, they didn't have a single Starlink satellite nor Starlink sub.
05:03Today, they have 10 million subscribers across the globe.
05:06Obviously, there are ambitions to do more on the wireless carrier front.
05:12So SpaceX has a lot going for it.
05:14And, of course, you know, the cherry on top would be the future of orbital data centers.
05:19Incredible capital efficiency, right?
05:21$12 billion raised in its lifetime.
05:23And you think about what it's done with that.
05:26The next thing is, what's a scenario where SpaceX merges with Tesla and why?
05:33Because that's the reaction that people have had since TerraFab was announced over the weekend.
05:39Yeah, I mean, obviously, people have been talking about that possibility for a little while.
05:44And SpaceX acquiring XAI certainly catalyzed those conversations even further.
05:50And I think Elon is, you know, kind of putting his cards on the table, if you will, as it
05:55relates to his ambitions to vertically integrate the ability to be a chip manufacturer, as difficult as it may be,
06:04but also in putting in aggregate data centers.
06:06Remember, he's done it here on Earth, Colossus 1 and Colossus 2.
06:10He did in record time, Colossus 1 specifically, and very large GPU clusters.
06:18And so from an engineering perspective, I think he knows how to do it.
06:21I'm not trying to be flippant about how difficult some of these things are, but he is an expert at
06:28vertically integrating companies.
Comments