Jan 23, 2025 · 1h 11m · bg2-pod
Stargate, Executive Orders, TikTok, DOGE, Public Valuations | BG2 w/ Bill Gurley & Brad Gerstner · Bg2 Pod
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 BG², hosts Brad Gerstner and Bill Gurley analyze the financial and technical reality behind the $500 billion Stargate AI announcement, evaluate shifting post-inauguration technology policies, and examine public market dynamics alongside guest ARM CEO Rene Haas.
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 85.6% of the talking time here. How this is scored →
Brad and Bill as informed peer 7.4 Guest teaching 0.9 Guest disagreement 0.0 Brad and Bill pushing back 2.0
speaking balance: gold is Brad and Bill, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 39:56 Haas reframes the Stargate entity structure
Rene Haas directly dismisses the notion that Stargate is an independent commercial business competing for outside workloads, clarifying it is strictly an operational vehicle for OpenAI.
Hardest push from Brad and Bill ▶ 39:56 Gurley presses Haas on lack of structural precedentBill Gurley repeatedly refuses to accept vague descriptions of Stargate's corporate form, pressing Haas on how an entity with 3:1 debt leverage can function without clear independent equity value.
Biggest teaching moment ▶ 42:15 Haas educates on non-obvious supply chain and deployment bottlenecksARM CEO Rene Haas systematically lays out real-world physical constraints that financial models miss, including TSMC 2nm/3nm wafer allocation, HBM bottlenecks, multi-site network latency, and physical cabling labor.
Brad and Bill hold their own ▶ 9:00 Gerstner's bottoms-up GPU financial breakdownBrad Gerstner dismantles sensationalist claims about Stargate's immediate cash demands by presenting precise unit economics showing a realistic initial $3B equity check for 250k GPUs.
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 |
|---|---|---|---|---|---|---|
| Host Welcome & Overview of Post-Inauguration Developments | 7 | 0 | 0 | 1 | Bill Gurley breaks down the White House Stargate announcement details, identifying participating entities (SoftBank, OpenAI, Oracle, MGX) and noting key details missing from the initial PR release. | |
| Compute Demand, Abilene Texas Facility & Financial Modeling | 7 | 0 | 0 | 2 | Brad Gerstner details Altimeter's bottoms-up financial modeling of data center buildouts in Abilene, Texas, while Gurley playfully pushes back on Brad's repetition of Abilene. | |
| Capital Expenditure & Financial Breakdown of Stargate Project | 8 | 0 | 0 | 2 | Gerstner provides a detailed breakdown of 2025-2026 CapEx and equity check requirements for 250k to 2M GPUs, challenging the media narrative that $500B cash is needed upfront. | |
| Multi-Year CapEx Scaling & Power Grid Deregulation (2026-2028) | 8 | 0 | 0 | 3 | Gurley challenges the assumed 3:1 debt-to-equity leverage ratio by comparing it against Equinix's historical 1:1 ratio, while Gerstner explains the energy requirements reaching 7.5 GW. | |
| Alternative Cost Analysis & Dylan Patel's SemiAnalysis Projections | 7 | 0 | 0 | 1 | Gerstner integrates Dylan Patel's TCO framework from SemiAnalysis, reconciling how Elon Musk and Sam Altman can both be technically correct regarding the $500B figure. | |
| Strategic Motivations & Competitive Dynamics of Key Tech Players | 8 | 0 | 0 | 2 | Gurley analyzes the corporate motivations of OpenAI, Microsoft, SoftBank, and Oracle, noting that Microsoft's reluctance to build OpenAI's custom mega-cluster drove Altman to seek outside partners. | |
| The Sport of Kings: AI Arms Race & Model Evolution | 8 | 0 | 0 | 1 | The hosts examine the massive CapEx escalating the AI race into a 'sport of kings' and discuss Chinese AI efficiency breakthroughs with DeepSeek models circumventing US chip sanctions. | |
| ARM CEO Rene Haas Joins to Discuss Stargate Role | 6 | 6 | 0 | 1 | ARM CEO Rene Haas enters to provide insider context on Masa Son's singularity vision and ARM's CPU role inside the Grace Blackwell GB200 architecture across Stargate. | |
| Stargate Governance, Fab Capacity & AI Supply Chain Bottlenecks | 7 | 7 | 0 | 4 | Gurley pushes Haas on the Stargate corporate governance structure and historical analogies, and Haas educates on physical constraints including advanced fab capacity (2nm/3nm), HBM, and human cabling labor. | |
| Host Debrief on Rene Haas & DeepSeek Chinese AI Breakthroughs | 7 | 0 | 0 | 2 | Following Haas's departure, Gurley expresses skepticism about Stargate's corporate governance clarity and explains his thesis that US export restrictions forced Chinese developers to innovate efficiently. | |
| Executive Orders, AI Diffusion Rule & Federal Preemption Strategy | 8 | 0 | 0 | 2 | Gerstner and Gurley dissect executive orders, the Commerce Department's AI Diffusion rule, and Congressman Jay Obernolte's legislative strategy for federal preemption over fragmented state AI regulations. | |
| DOGE Execution Strategy & Path to Balancing U.S. Budget | 7 | 0 | 0 | 3 | Gurley offers a sober perspective on the probability of DOGE overcoming entrenched federal bureaucracy, while Gerstner models potential interest rate relief from a balanced budget commitment. | |
| Trump's TikTok Strategy & U.S. Government Equity Stake Proposal | 7 | 0 | 0 | 3 | Gurley pushes back on Trump's proposed TikTok equity stake scheme, highlighting that China's government must consent and noting that existing statutory deadlines limit executive discretion. | |
| Public Markets, Stan Druckenmiller Insights & Tech Earnings Productivity | 8 | 0 | 0 | 1 | Gerstner analyzes public market multiples, Stan Druckenmiller's CNBC macro commentary, 10-year Treasury yields, and Netflix's earnings beats driven by AI productivity gains. |
Statements from this episode (30)
Gerstner: Trump, Ellison, Altman, and Son Made White House Stargate Announcement
“There was a huge announcement at the White House yesterday with Larry Ellison, Sam Altman, Massa Sun, and President Trump.”
Gurley: Microsoft leads tech capex at roughly $80 billion per year
“I guess Microsoft's the highest right now at eighty billion a year.”
Gurley: Stargate sounds like a better-funded version of CoreWeave
“It sounds to me like maybe a, an even better funded version of CoreWeave”
Gerstner: OpenAI Is Already Running Workloads Out of Abilene Facility
“Larry said at the press conference, They have 10 buildings already built in Abilene. They have another 10 under construction, and we know that OpenAI is already running workloads out of Abilene, Texas, which is where this mega project's going to be built, and …”
Gerstner: Stargate's 2025 Buildout Needs Only ~$3B in Equity
“Out of the thirteen billion dollars of capex required, you know, if you make an assumption about 25 75 or 30 70 equity to debt, that probably requires like a three billion dollar equity check.”
Gerstner: Deploying two million GPUs requires about 2.7 gigawatts of power
“To stand up two billion or two million GPUs in 2026, you need about 2.7 gigs, gigawatts of power, right?”
Gerstner: Energy deregulation is probably the Trump administration's top priority
“If there's anything that I believe to be true, drill baby drill, and tapping US energy reserves, and removing obstacles and regulations in order to light all this up in America, rather than other places, is probably priority number one.”
Gerstner: Trump expanded Stargate project scale from $100B to $200B at Mar-a-Lago
“Trump, when they were down in Mar-a-Lago, took it from a hundred billion to two hundred billion, and now they got to five hundred billion, which served everybody around the table really well.”
Gerstner: No entity has $100B or $500B upfront for Stargate
“Elon is 100% right. Nobody has five hundred billion dollars. In fact, nobody has a hundred billion dollars to contribute to this on day one.”
Gurley: Microsoft refused to build massive AI cluster Altman wanted
“I think my own interpretation is Microsoft didn't want to build this big, you know, open AI playground or one as big as maybe Sam wanted to build. And so he went somewhere else and everybody, But everyone at the surface behaving like they're okay with that.”
Gerstner: xAI will launch Grok 3 trained on world's biggest cluster
“We're going to see he's going to launch Grok three here. I think they're going to have the output the first training run on the biggest cluster in the world. And we're excited to see what that yields from my own perspective.”
Gurley: Nvidia rivals cannot ramp chip production fast enough
“Even if an NVIDIA competitor showed up, I don't think they can create enough production in the amount of time you need.”
Gerstner: Anthropic has under $1B revenue and lacks rival traction
“You know, is Anthropic going to be able to show, you know, they have less than a billion in revenue. They don't really have they have good, great models, I would argue, but they don't have the consumer or the enterprise traction that the other folks do.”
Gerstner: Chip and power bottlenecks will force subscale AI mergers
“I think you are going to be chip constrained. I think you are going to be power constrained. You know, for the next several years, and I think if you're subscale, On any level, I think those players ultimately have to get folded into a much larger entity.”
Gurley: US policy to block China from AI is futile and backfired
“I think the policy of American, of the American government to try and keep China out of the AI game is futile. I've always felt that way. I don't think it'll work. In this case, I actually think it backfired.”
Gerstner Forecasts 250k GPUs in 2025 and 1-2M in 2026 for Stargate
“And I'm forecasting maybe 250,000 GPUs this year, maybe a million or two million, you know, next year”
Haas: Energy Needs for AGI Exceed Big Tech Balance Sheets
“Get access to so much energy that's needed, we think, to drive AGI and ASI at numbers that are even well beyond the balance sheets of the giant Companies like a Microsoft, a Google, a Meta, an AWS, et cetera, an Amazon.”
Haas: The Base CPU for Stargate AI Data Centers Will Be ARM
“And every data center whether it's running general purpose compute, whether it's running inference or running training, needs a base CPU to run everything, end quote. And that's our role. And whether that is going to be what we are part of today, which is GB 2…”
Haas: OpenAI will have operational control of the Stargate project
“And the operational control will be from OpenAI. So they'll, they'll call the shots relative to all the things relative to the operation.”
Haas: TSMC leading-edge capacity and HBM will constrain AI buildout
“Fab capacity does become an issue. No, no doubt. Because when you start thinking about three nanometer and two nanometer you know, TSMC is, is the leader, is the, Only game on town on some level. So I think that is a potential limitation or at least constraint…”
Gerstner: Commerce Department AI diffusion rule will slow down American AI
“There's a very insidious piece regulation, again, under the guise of being tough on China that I think is really gonna slow down American AI, and this is a framework for artificial intelligent diffusion, and it was promulgated by the Department of Commerce, an…”
Gurley: State-by-state AI regulation is the dumbest possible approach
“Because the dumbest possible thing we could possibly have is state by state. It just, like, it's just mud in the gears, like really, Really stupid.”
Gerstner: Congress plans federal preemption and moratoriums on state AI laws
“I coordinated some meetings this weekend between him and a lot of CEOs of leading AI companies. And there is a plan by Congress on the one hand to have federal preemption. Clearly this is a matter of interstate commerce. So we should have federal preemption, b…”
Gurley: DOGE achieving radical federal efficiency is high impact but low probability
“I put it at high, high impact, still low probability in my brain, just because it's such a bureaucratic place, Washington”
Gerstner: Trump will commit to balancing budget in March State of the Union
“I think in the President's State of the Union address in the first week of March, he's gonna make a commitment to balance the budget in his first term in office.”
Gurley: Banning TikTok Would Shift Most of Its Value to Meta
“If this law is, is enforced, goes away. And most of that equity value probably goes to meta, right? And maybe a little bit to snap or if someone else pops up, but it just melts away.”
Gurley: Odds of ByteDance Agreeing to Sell TikTok Are Low
“Well, and they've had a 180 days to do it. So the odds that they actually want to do it, I think are low because they could have done it already.”
Gerstner: S&P excluding Mag 7 faces high hurdle from baked-in earnings acceleration
“If you look at the S and P, we think that S and P is baking in a big earnings acceleration this year. And that's on top of already high multiples. So the S and P X mag seven has a big hurdle this year. If you look at what's expected out of mag seven, it's a pr…”
Gerstner: Altimeter will reduce exposure if interest rates reach 5% to 5.5%
“If rates were to go to five or five and a half percent and we see that happening, our exposures would come down”
Gerstner: AI productivity will drive a five-year golden era of margin expansion
“I've said that I think this moment that we're in the next five years is going to be a golden moment of margin expansion for technology businesses, and frankly, all businesses. Human productivity is going to go up, which means the cost to do exactly, you know, …”