Jun 11, 2026 · 1h 20m · bg2-pod
The SpaceX IPO, Fable 5, AI Capex Update & Market Check w/ Gavin Baker, Andrew Fox & Clark Tang · Bg2 Pod
Brad Gerstner · 33m spoken Gavin Baker · 22m spoken Clark Tang · 9m spoken Andrew Fox · 6m spoken Jensen Huang · 24s spoken
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the BG² podcast, Brad Gerstner, Gavin Baker, Andrew Fox, and Clark Tang evaluate SpaceX's projected $1.77 trillion IPO, the convergence of aerospace and AI compute infrastructure, Anthropic's Fable 5 release, and macro market positioning.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Brad and Bill hold 45.8% of the talking time here. How this is scored →
Brad and Bill as informed peer 6.6 Guest teaching 3.8 Guest disagreement 1.2 Brad and Bill pushing back 1.4
speaking balance: gold is Brad and Bill, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 56:53 Gavin mocks custom ASIC threats against Nvidia
Gavin sarcastically dismisses hyperscaler in-house ASIC efforts ('cute ASIC you've built there') and explains how Nvidia could retaliate by releasing frontier open-source models.
Hardest push from Brad and Bill ▶ 53:09 Brad rejects the open-source commoditization thesisBrad directly challenges the consensus argument from earlier shows, citing empirical data that frontier models capture over 90% of actual market revenue.
Biggest teaching moment ▶ 26:40 Fox details space compute unit economicsAndrew Fox educates the room on the precise mechanics of Starship payload economics, proving how orbital compute achieves a 5x CapEx reduction per gigawatt over terrestrial data centers.
Brad and Bill hold their own ▶ 1:03:36 Brad contextualizes global AI CapEx sustainabilityBrad synthesizes hyperscaler capital raises, Morgan Stanley forecasts, and inference gross margins to demonstrate exactly why a $1.5T CapEx cycle is economically viable.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Brad and Bill as informed peer | Guest teaching | Guest disagreement | Brad and Bill pushing back | Why |
|---|---|---|---|---|---|---|
| BG² Podcast Title Sequence & Visual Montage | 6 | 2 | 1 | 1 | Brad opens by establishing detailed IPO metrics ($135/share, $1.77T valuation, $160B 2028 revenue projections) and framing the key debate drivers. Gavin acknowledges Altimeter's proprietary data and charts on Colossus IRR and operating profits per gigawatt. | |
| SpaceX IPO: Coding Agents, Cursor Benchmarks & AWS Parallel | 6 | 5 | 1 | 1 | Gavin explains how Composer 2.5 achieved Pareto dominance on CursorBench due to proprietary coding tokens, educating the room on model frontier shifts. Brad extends the insight by comparing SpaceX's compute infrastructure moves to Amazon's early AWS strategy. | |
| SpaceX Core Business: Launch Cadence & Starlink Economics | 6 | 4 | 2 | 2 | Fox lays out launch cadences and Starlink economics, while Gavin interjects with realistic skepticism about the difficulty of rapid full reusability. Brad pushes for concrete probability distributions and models the $10B to $50B connectivity growth curve. | |
| Elon Web Services (EWS) & Hyperscale Compute Buildout | 7 | 3 | 1 | 1 | Brad and Clark analyze Elon Web Services (EWS) and cite Jensen Huang's 19-day cluster bring-up metrics. Gavin reinforces that custom behind-the-meter data center engineering is a non-commodity moat generating top-tier monetization. | |
| Space Data Centers & Orbital Compute Unit Economics | 6 | 6 | 1 | 2 | Fox and Gavin detail the unit economics of space data centers, calculating $5B/GW orbital CapEx versus $25B terrestrial facility costs. Brad presses on whether orbital compute is an immediate necessity for the IPO or merely an upside call option. | |
| X.AI's Cursor Acquisition & Frontier Model Upside | 7 | 4 | 1 | 2 | Brad argues that xAI's acquisition of Cursor and its internal model potential is the most overlooked upside driver for SpaceX. Gavin and Clark validate the thesis, pointing to Grok 4.3's frontier placement and Vera Rubin chip allocations. | |
| SpaceX IPO Bull vs. Bear Case & Post-IPO Trading Dynamics | 7 | 3 | 1 | 1 | Brad reviews historic post-IPO drawdowns and explains Altimeter's ballast portfolio strategy. Gavin strongly agrees, noting the unprecedented liquidity history of SpaceX private tenders and the 365-day insider lockup structure. | |
| Anthropic Fable 5, Mythos & Test-Time Compute Paradigm | 7 | 5 | 2 | 2 | Gavin highlights Noam Brown's insights on test-time compute and long-running models outsmarting fixed benchmarks, while Clark shares firsthand testing of multi-agent orchestration. Brad notes this decisively disproves the narrative that open source commoditizes frontier intelligence. | |
| Custom ASICs, NVIDIA Dominance & Global AI CapEx Validation | 8 | 5 | 2 | 2 | Clark and Gavin dissect custom ASICs versus Nvidia's token-per-watt efficiency lead. Brad delivers a comprehensive analysis balancing $1.5T in projected 2027 CapEx against hundreds of billions in rising inference revenue. | |
| Market Check & Altimeter Exposure Adjustment | 7 | 1 | 0 | 0 | Brad delivers a market check monologue detailing why Altimeter shifted exposure from large to medium-small after a massive semiconductor run, warning about macro headwinds and inflation overhangs. | |
| Gavin Baker's Macro Outlook & AI Token Seasonality | 6 | 4 | 1 | 1 | Gavin uses the analogy of a runner reaching a cliff top to describe overextended tech valuations, explaining historical summer token seasonality while maintaining an aggressive long-term bullish stance on AI capabilities. |
Statements from this episode (36)
Gerstner: Altimeter holds SpaceX shares and expects to buy in IPO
“We're also a shareholder.
you know, we also expect to be buying in the IPO.”
Baker: xAI's Google deal yields higher profit per GW than competitors
“He shows that XAI's deal with Google for cloud computing generates more operating profit per gigawatt than Anthropic, than Meta, than Google, than OpenAI.”
Baker: Musk builds data centers faster than anyone at 122 days
“We do know from Jensen that
Elon brings data centers up faster than anyone, a 122 days.”
Baker: All frontier AI model revenue will accrue to the Pareto curve
“And I do think being all revenue will accrue to the Pareto curve. All at least kind of frontier model revenue will accrue to the Pareto curve.”
Baker: Cursor and Anthropic hold more coding data than public web
“But my understanding is that Cursor and Anthropic have more tokens of proprietary coding data than anyone else, and they each have more tokens of proprietary coding data than exist on the public internet.”
Fox: SpaceX Launch Cadence Will Reach Thousands of Flights
“So look, I think expectations for now, you know, we're going from, You know, call it a 160, 165 launches last year, up into the high hundreds of launches in the next several years. And, you know, getting into the thousands of launches probably in the next thre…”
Fox: Starlink Penetrates Under 1% of Global Households
“You look at the percent of global households with Starlink, it's less than one percent.”
Baker: $50B in Starlink Revenue Is 0.3% of Global Telecom TAM
“Fifty billion, that's, you know, 0.3% penetration of the global telecom market.”
Huang: xAI Brought a 100k GPU Cluster Online in 19 Days
“Just to put in perspective, a 100,000 GPUs, that's You know, easily the fastest supercomputer on the planet as one cluster. A supercomputer, ah, that you would build would take normally three years to plan. And then they deliver the equipment, and it takes one…”
Gerstner: Musk May Emerge as Largest Player in AI Compute
“So I think there might be a real consolidation in the AI compute market where you have the hyperscalers on the one hand, and on the other hand, you know, he may emerge as the largest, strongest player in the AI compute market.”
Baker: Musk's Infrastructure Became the Fourth-Largest AI Hyperscaler
“In 30 days, we went from not being an AI hyperscaler to being number four, and we passed a lot of companies, including Oracle.”
Fox: xAI signed compute deals with Anthropic and Google worth billions
“They just signed Anthropic at 22 to 23. They just signed Google at 50.”
Fox: Starship Can Carry 5 Megawatts of AI Compute Per Launch
“And when you back into the numbers, you get to something like five megawatts of capacity per Starship launch.”
Fox: Orbital Compute CapEx Is $5B/GW vs $20-25B Terrestrially
“The math you get to before you account for things like bad GPUs, bad satellites, right, these will all be things that happen, but the math you get to is it's about five billion dollars per gigawatt Of CapEx to put these in space. Right. For comparison, terrest…”
Baker: Ground AI Data Centers Cost $60B Per Gigawatt Today
“It costs sixty billion dollars to put a gigawatt on the ground today. Yeah. And we'll call it 35 of that is, are the GPUs and the silicon that's doing the training and the inference, and twenty-five billion is the land, the shell, the power, and the cooling.”
Gerstner: Altimeter projected Cursor could exit the year at $10B revenue
“Our own, ah, projections were that they could exit this year at up to ten billion dollars of revenue, so they were growing very fast, ah, ah, one of the leading coding agents.”
Baker: Grok 4.5T parameter model is injecting Cursor data into pre-training
“The Grok 4.5 trillion parameter model is training. One would hypothesize based on scaling laws that that will might be a better base model. And then the cursor data is being injected into the pre-training process, not just reinforcement learning.”
Baker: Grok 4.3 was the most intelligent 500B parameter model in the world
“Grok 4.3 was on the Pareto frontier, and as of 10 or 12 days ago, and these things move fast, most intelligent, five hundred billion parameter model in the world. And they were on the frontier, and there are four companies on the frontier. XAI, SpaceX AI, Goog…”
Gerstner: SpaceX IPO is a must-own for institutional investors
“And so I think for most institutional investors, it's a must buy, a must own, it's set it and forget it, right, in order to have a real bet on both the space and the AI future.”
Gerstner: Elon Musk is locked up for 365 days post-SpaceX IPO
“And by the way, he's locked up for 365 days. Yes, exactly. So we know he's not selling, right?”
Baker: SpaceX has offered employee liquidity every six months for 10 years
“I think another thing that's unprecedented here is the employees and to a large degree the investors here have had liquidity every six months. For like the last 10 years. So if you're a SpaceX employee or former employee and you wanted to sell, you've had, wha…”
Baker: Frontier AI Will Capture Most Revenue While Open Source Drives Token Volume
“The majority of economic value may continue to accrue to the frontier, and man has it ever accrued to the frontier thus far, and for sure the first six months this year, but the majority of tokens consumed in the world may be open source.”
Baker: Open-Source AI Growth Directly Benefits Hardware and Compute Providers
“It's actually really bullish for compute and hardware, because if the frontier models are capturing less of the margin, then you're gonna spend more on compute. So the better open source does, the better it is for compute providers.”
Baker: NVIDIA is highly likely to dominate open-source AI
“I do think NVIDIA is highly likely to be the world's dominant provider of open source AI, and I do think Jensen will bring open source, you know, right now it's, whatever, six months behind the frontier. We might see it creep closer and closer and closer”
Baker: NVIDIA can become a top cloud computing company faster than expected
“I think Nvidia can join the frontier and become one of the world's largest cloud computing companies. Much faster than people think.”
Gerstner: Global AI CapEx will approach $1.5 trillion in 2027
“I think the number on twenty-twenty-seven Is likely closer to 1.5 trillion.”
Baker: Annual AI inference revenue will finish the year well over $200B
“I think we end this year well over two hundred billion in inference revenue, well over.”
Tang: AI monetization per gigawatt rose from $20B to $30B–$40B
“When the monetization per gigawatt is actually increasing from, you know, call it, like, 20, twenty billion in the best, best of cases for, at the beginning of the year, To now, like, 30 to even pushing 40.”
Tang: AI compute supply will fall far short within three years
“If we continue this curve within like three years, you know, we're just gonna be so short on all the compute.”
Gerstner: Anthropic achieved accidental profitability due to compute spending constraints
“Atropic, no way did they think they were going to be anywhere close to breakeven. Right? In, in, in, in this part of the curve. And the reason, like I called it accidental profitability that, you know, people have been talking about that because they want to s…”
Gerstner: Software and Internet Lagged While AI Lifted SPY and NASDAQ
“Internet's down 16%. Software's down eight percent on the year. You know, SPY and NASDAQ are up, but really up because of their components that are related to AI and compute.”
Gerstner: The Whole Market Could Be Down Without Anthropic's Revenue
“I think, you know, I've said it a couple times, I think if the anthropic revenue had not shown up this year, because that was the overhang on the market, I think the whole market could be down this year, right?”
Baker: Finding the next AI infrastructure bottleneck is an outdated trade
“Like I do see a lot on X about finding the next bottleneck. I think that was the last game. That game is over.”
Baker: AI token consumption slows during summers due to student seasonality
“AI has actually been seasonal for the last three summers. Token consumption is kind of plateaued, slowed down, and that's because, you know, college kids are big AI consumers, and they don't use as much AI, you know, hopefully they're all using it to learn and…”
Gerstner: Mag-Seven added $1T revenue and $17T market cap in 7 years
“In the last seven years, we've added one trillion of revenue to the Mag-Seven in the last seven years, ok? To get to a trillion, to get to the first trillion Of, you know, took over 20 years, and the last seven we added another trillion, and that added 17 tril…”
Gerstner: SpaceX, Anthropic, OpenAI will add $1T revenue in 4-5 years
“The forecast now that we're going to add another trillion of revenue in just three companies, SpaceX, Anthropic, and OpenAI, over the next four to five years, ok? Like, not seven companies, three companies, and in half the time, right?”