May 22, 2026 · 1h 42m
SpaceX targets two trillion valuation as Nvidia faces shock stock selloff
SpaceX's $2T Case, Nvidia's Shock Selloff, America Turns on AI, Trump Pulls AI Order, Bond Crisis?
The intersection of SpaceX's historic private valuation, Nvidia's market volatility, and shifting federal AI policy will redefine the next decade of technology investing.
- 1SpaceX's recent S-1 filing outlines a financial trajectory that could push its valuation to two trillion dollars.
- 2Nvidia's record earnings triggered a surprise stock selloff, signaling market caution over heavy AI infrastructure spending.
- 3Donald Trump's decision to rescind the AI Executive Order marks a major shift in national technology policy.
Don't miss
The panel debates the financial disclosures in SpaceX's S-1 filing and what it means for the future of private space flight.
The brief
The All-In crew is joined by investor Gavin Henry to dissect SpaceX's potential path to a two trillion dollar valuation, following the massive implications of the aerospace giant's recent S-1 filing.
While Nvidia posted record-breaking earnings, a sudden stock selloff highlights growing market anxiety over the massive capital expenditures poured into artificial intelligence infrastructure.
Public sentiment toward AI is shifting as geopolitical strategies realign, highlighted by Donald Trump's decision to pull the current AI Executive Order to reshape national tech policy.
What was said on this episode
35 statements · 27 positive · 5 negative · 1 mixed · 2 neutral
Recursive AI could improve model quality tenfold annually.
“you can potentially live out this idea that there's an order of magnitude improvement on a yearly basis”
Listen at 4:08
OpenAI and Anthropic may have $100 billion ARR with roughly 80% inference margins.
“if OpenAI and Anthropic are at, call it $100 billion of ARR now with 80% ish gross margins on inference, like the returns are there”
Listen at 5:00
AI language-model companies could reach $200–400 billion ARR by year-end.
“it's, it's not hard to see 200, 300, $400 billion of ARR the end of this year at high margin”
Listen at 5:20
Continual learning would let AI models learn from experience like humans.
“continual learning is the holy grail, where the model learns from experiences the way humans do”
Listen at 6:55
A minor architectural breakthrough could halve AI inference cost per token.
“All it takes is a minor breakthrough and your cost per token drops in half.”
Listen at 8:48
Dario Amodei has an incentive to promote regulation creating a moat for Anthropic.
“It makes enormous sense for Dario to try to create the boundary conditions for a regulatory moat”
Listen at 13:42
AI may initially produce outsized returns for its controllers relative to society.
“a small number of people that control and profit from and benefit from AI are going to end up getting outsized returns relative to the broader population”
Listen at 19:19
If the United States slows AI advancement, another country will advance instead.
“if the United States does not advance its AI technology, the availability of it, tbd, industry, taxation, all these things that we're talking about doing, there will be someone else that will”
Listen at 24:17
The United States and China should apply KYC controls to advanced AI models.
“we should probably do some sort of KYC so some crazy person doesn't create some biological weapon”
Listen at 29:09
Existing courts provide recourse against irresponsible AI model companies.
“if an AI model company behaves irresponsibly, they know that they're are ways that people who have been harmed can seek recourse”
Listen at 31:08
Cities without autonomous ride services will feel unsafe and backward.
“I think going to a city where you can't get an awaymo or a cyber cab is going to feel barbaric and unsafe”
Listen at 32:48
Restrictions preventing autonomous vehicles will not persist because of convenience benefits.
“from a convenience and quality of life perspective, I just don't think it's going to persist”
Listen at 33:28
Flock Safety technology can substantially solve crime if municipalities choose it.
“We can really, really solve crime, and it's just your choice.”
Listen at 35:14
An annual $30–40 million investment could make Las Vegas America’s safest city.
“If you gave the Las Vegas Police Department 30, $40 million a year, it would be the safest city in America, and that's all it would take.”
Listen at 36:33
SpaceX builds data centers faster and more cheaply than competitors.
“They build data centers dramatically faster than anyone else at a lower cost.”
Listen at 48:28
SpaceX’s AI infrastructure business could grow far faster than previously expected.
“this business can grow dramatically faster than I think maybe what anyone could have contemplated three months ago”
Listen at 49:07
Cursor’s Composer 2.5 became Pareto-dominant after three weeks of reinforcement learning.
“Composer 2.5 is literally well outside the Pareto frontier. And that's after three weeks.”
Listen at 51:10
Major technology companies believe orbital computing will become real.
“Google, Anthropic, Amazon, Nvidia, they are all very convinced that Orbital Compute is going to be reality.”
Listen at 52:40
Agentic AI harnesses are approximately as important as the underlying models.
“the harness is essentially as important as the model, especially in an agentic world”
Listen at 54:25
Space-based communications and data centers would provide a valuable alternative infrastructure layer.
“having like a space based communication network, space based data centers and space based communication back down to Earth wireless, I think is generally a good thing”
Listen at 57:40
SpaceX revenue could reach $40–45 billion next year and double the following year.
“I suspect what happens is next year it's probably 40, 45 billion, and then the year after that it probably doubles again.”
Listen at 1:01:44
SpaceX terrestrial data centers could generate $100–200 billion revenue by 2030–2032.
“terrestrial data centers alone are 100 or $200 billion of revenue by 2030. 2032.”
Listen at 1:03:34
Rapidly reusable rockets can launch and land multiple times daily.
“Rapid reusability means that you can fly the same fly and land the same rocket. Multiple times per day.”
Listen at 1:07:35
SpaceX may achieve rapid Starship reusability within one or two years.
“I would guess a year or two, maybe sooner.”
Listen at 1:09:10
Orbital computing will become commercially relevant between late 2028 and early 2030.
“second half of 28 to first half of 2030 would be my point prediction”
Listen at 1:11:12
AI-related asset valuations are currently inconsistent across market segments.
“The AI market is cross sectionally inefficient right now.”
Listen at 1:15:18
Nvidia is gaining AI semiconductor share despite claims of share loss.
“How could there be a share loss narrative if I am gaining share?”
Listen at 1:17:15
Nvidia expects its CPU business to generate $20 billion this year.
“they said if they thought their CPU business was going to be $20 billion this year.”
Listen at 1:18:28
Older GPUs can remain useful for 10–15 years in disaggregated accelerator systems.
“those older GPUs, they have a useful life for 10 or 15 years”
Listen at 1:21:11
High Japanese yields and carry-trade exposure could trigger a credit crisis.
“I would think that this is one of those things that could be a catalyst for a, for a credit crisis”
Listen at 1:25:24
Investors should limit their actively tracked public stocks to five or fewer.
“It's five. It's five or less.”
Listen at 1:27:06
Nvidia trades at roughly a low-to-mid-teens multiple of real earnings.
“I think Nvidia is probably at a low teens multiple of kind of low to mid teens multiple of real earnings.”
Listen at 1:28:56
A prolonged Strait of Hormuz closure benefits America relative to Europe and Asia.
“the street of Hormuz is absolutely bad for everyone, but relatively good for America and relatively good for Trump's policy goals”
Listen at 1:31:06
The United States–China meeting did not produce the hoped-for grand de-escalation deal.
“Fundamentally the grand deal, the big deal that I would say reduces de escalates tension, probably didn't manifest as some had hoped it would.”
Listen at 1:34:10
Selling older Nvidia GPUs to China reduces China’s incentive to build an alternative ecosystem.
“selling and deprecated Nvidia GPUs to China lowers the odds of them developing their own alternative ecosystem”
Listen at 1:36:19
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.