Feb 22, 2024 · 1h 21m · bg2-pod
BG2 w/ Bill Gurley & Brad Gerstner | NVDA, Chips, AI Compute Build Out, AI Impact on Big Tech | E03 · 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 episode 03 of the BG² podcast, venture capitalists Brad Gerstner and Bill Gurley analyze how artificial intelligence is reshaping public tech stock valuations, Big Tech competitive moats, and consumer product architecture. They detail the transition toward contextual memory and autonomous action models while evaluating the multi-trillion-dollar global compute, power, and semiconductor buildout.
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 99.7% of the talking time here. How this is scored →
Brad and Bill as informed peer 7.5 Guest teaching 2.1 Guest disagreement 2.3 Brad and Bill pushing back 3.1
speaking balance: gold is Brad and Bill, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 40:38 Direct pushback on UI screen-scraping vs APIs
Bill bluntly rejects Brad's praise of Large Action Models controlling screen pixels, calling it an inefficient hack compared to structured service APIs.
Hardest push from Brad and Bill ▶ 1:01:44 Skepticism of fab startup and multi-trillion fundraising hypeBill dismisses the notion that startups can outbuild TSMC or Nvidia simultaneously, assigning nearly zero mathematical probability to such ambitious ventures.
Biggest teaching moment ▶ 1:04:50 Morris Chang analysis on semiconductor manufacturing economicsBill educates listeners on the cultural and operational labor realities highlighted by TSMC founder Morris Chang, explaining why US fab re-onshoring faces severe workforce hurdles.
Brad and Bill hold their own ▶ 18:50 Granular unit economic analysis of AI search disruptionBrad demonstrates deep analytical mastery by comparing the fractional-cent cost of traditional 10 blue links against the 10x-50x inference cost of generative responses.
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 |
|---|---|---|---|---|---|---|
| In-Person Recording Format and Podcast Dynamics | 5 | 0 | 1 | 1 | Brad and Bill open the episode with light banter about FSD v12 and discuss how the intellectual rigor of preparing for the podcast forces deep research. | |
| Market Backdrop, Tech Dispersion, and Inflation Signals | 7 | 2 | 2 | 3 | Brad breaks down macro indicators including hot CPI prints, real interest rates at 2007 levels, and Larry Summers' hawkish warnings while Bill probes the certainty of a soft landing. | |
| Long-Term Conviction vs. Short-Term Trading and CAP Framework | 8 | 2 | 1 | 2 | Bill introduces Michael Mauboussin's Competitive Advantage Period (CAP) framework, connecting it to historical tech valuation traps like BlackBerry and Yahoo alongside Brad's 120-bagger Booking.com case study. | |
| Big Tech Analysis: Microsoft's Moat vs. Google Search Disruption | 8 | 2 | 2 | 3 | Brad details the unit economics threatening Google Search, explaining how shifting from ten blue links (fraction of a cent) to multi-token answer generation (4+ cents) destroys search gross margins from 95% down to 50%. | |
| Value Capture Phase Shifts and AI Memory Breakthroughs | 8 | 2 | 2 | 3 | The hosts debate consumer AI business models and technical hurdles around persistent user memory, with Bill noting that personal context would require daily fine-tuning models. | |
| MANG Stock Valuations and Large Action Models | 7 | 2 | 2 | 2 | Brad presents valuation metrics across the MANG group highlighting growth-adjusted PEG ratios, while Bill emphasizes the switching costs and moats established once personal memory is locked into a platform. | |
| AI Execution Architectures, APIs, and Ecosystem Dynamics | 8 | 3 | 4 | 5 | Bill directly challenges Brad's enthusiasm for Large Action Models and pixel-scraping UI automation, arguing that direct API integrations will prove vastly superior. | |
| Big Tech Deep Dive: Meta Rebound and Apple Strategy | 7 | 2 | 3 | 4 | Brad explains Meta's operational turnaround through open-source AI, GPU infrastructure investments, and WhatsApp monetization, while Bill challenges whether Meta hardware can compete with Apple or Google's mobile distribution. | |
| Big Tech Deep Dive: Amazon E-Commerce, AWS, and Inference | 7 | 2 | 2 | 3 | The conversation covers Amazon's data gravity preserving AWS market share against Azure, the rising competition from Temu's AI supply chain, and custom silicon tradeoffs between training and inference. | |
| Sovereign AI Compute Buildout and Semiconductor Fab Geopolitics | 9 | 4 | 4 | 5 | Bill dismantles the feasibility of sovereign fab buildouts and multi-trillion fundraising plans, citing Morris Chang's MIT lecture to explain why semiconductor manufacturing cannot easily be re-onshored. | |
| Data Center TAM Expansion, Power Limits, and Market Cycles | 8 | 2 | 2 | 3 | Brad outlines the roadmap for data center capex doubling to $2 trillion, analyzing power limitations and sovereign energy advantages before both hosts discuss impending cyclical pullbacks. |
Statements from this episode (29)
Gerstner: Tesla FSD v12 will be a ChatGPT moment for autonomous driving
“I think this is going to be a bit of a chat GPT moment for full self-driving, but what it really, it just reminded me of the magic of this moment.”
Gerstner: Tesla made more FSD progress in 12 months than past 7 years
“Tesla rebuilding their models, For how they do self-driving around imitation learning and all this interesting stuff going on over there. Like, you know, I think they probably made more progress in the last 12 months than in the last seven years.”
Gerstner: Tech stock gains in early 2024 are concentrated in Meta and Nvidia
“Meta, Nvidia up 30, 40% even with the pullback that we had today. But the truth is the Nasdaq hasn't really moved that much. I mean, I think we're up three or four percent through today. If you look at the median stock, I think it's up about one percent.”
Gerstner: US real interest rates are at their highest level since 2007
“Real rates, right? So real rates, this is the restrictiveness that we have in the economy. So this is effectively the interest rates we have less the expected inflation rate in the future. They're as high as they've been since the fall of 2007, and the last ti…”
Gerstner: Fears of an Impending Nvidia Chip Glut Are Inaccurate
“So everybody's just the wall of worry around NVIDIA is when does the glut come? We've pulled forward all this training demand. It's dark fiber like in the year 2000. I think those things are not accurate.”
Gurley: Microsoft is a clear net winner from generative AI
“People looked at the Microsoft portfolio of assets, especially where they make money around the office suite, right? And said, and the developer community where they also, you know, control a lot of the IDEs that are used to program. And they said, oh, this is…”
Gerstner: Google search query growth slowed to 4% per year
“The number of queries that are asked of Google has slowed down a lot, right? So they're growing at about four percent a year. Monetization is up a lot, 13%, more ads on the page. We've talked about that. And so you have a 17% CAGR around search.”
Gerstner: AI search query answers cost 10x more than 10 blue links
“So if you say, like, what's it cost to provide 10 blue links? It's about a third of a penny or less per query. Now, what does it cost to do that for 750 tokens today? And of course, this will go down over time, but it's 10 X more, right? It's four cents. Per q…”
Gerstner: Google search market share and gross margins will decline
“Your share likely goes down over time because you have people like ChatGPT you have to compete with. But worse yet, your margin on each of those queries goes down over time. I don't know what it nets out at, 50%, 60%. Now, mind you, for perplexity or co-pilot …”
Gerstner: AI answer engines will operate below margin to take search share
“It would be surprising to me if we don't see Meta AI, and Microsoft, and ByteDance, and Perplexity, and all the others who are providing answer engines, ChatGPT, coming at them below margin, ok?”
Gurley: No consumer AI answer engine has shown holistically superior answer quality
“I, as of right now, and I've played with all of them, on the consumer side, I don't think anyone has a, perplexity came out, it was a lot faster. That was pretty cool. But when I just look at the quality of the answers, I don't, I mean, on different searches, …”
Gerstner: Apple will use small on-device models and delegate complex queries
“Think about what Apple's going to do maybe more like a small language model. Like, really understand all the language, really understand everything about me, really understand how I interact with all these applications, and then when I have a deep problem I ne…”
Gerstner: Google trades at the lowest P/E and growth rate among MANG peers
“This is the 23 through 25 expected growth rates. 14% for Microsoft, 12% for Amazon. 42% for NVIDIA and 11% for Google. So Google's already expected out of those four to be growing at the slowest rate. But then what's interesting, if you come down here, To the …”
Gurley: AI memory systems create massive switching costs and lock-in
“One thing that drives durability, going back to our competitive advantage period, is switching costs. If I start relying on one of these things as my memory, and I don't have a way to pull that out and jump to something else, I'm stuck. Like, I am hook, lock, …”
Gerstner: Autonomous driving will shift entirely to imitation learning over rules
“Well, here's where I think the world's going. We met with these robotics companies. We meet with Tesla, et cetera. Imitation learning, okay? They're not even gonna know what the stop sign is, or the traffic light is, or the dog in the street. They're not gonna…”
Gurley: Apple and Google will block AI apps overlaying mobile OSs
“Or some of them might try and sit on top of your phone. Of course, Google and Apple will stop them from doing that.”
Gurley: Ecosystem partners distrust Google and prefer OpenAI and Perplexity
“I don't think Google has treated its partners well in the search ecosystem. And so there's a lot of angst there and a lot of mistrust. And so if, OpenAI or perplexity came along and said, would you integrate and pass tokens? They might say yes. I think they're…”
Gerstner: Meta completely recovered from Apple privacy changes using AI
“Basically, because of AI, they've been able to backfill that monetization completely. Nobody thought, no investors thought that was going to be the case 18 months ago. So their core business got a lot stronger.”
Gerstner: On-device AI models will unlock a huge Apple upgrade cycle
“To me, the big breakthrough on Apple is if they can run a five, ten billion dollar parameter model on the phone, on the edge, without consuming all of my battery, which, you know, there's a lot of talk that they're going to be able to do that. They can maintai…”
Gerstner: OpenAI has caused almost no share shift from AWS to Azure
“We've seen almost no share shift from AWS to Azure as a result of OpenAI.”
Gerstner: Amazon will deploy custom silicon but remain major Nvidia buyer
“I think the answer to that is yes, but I still think they'll be one of the largest buyers of NVIDIA GPUs in the world because it doesn't replace that for a lot of really important workloads.”
Gurley: AI inference market will prioritize power and cost over raw performance
“The inference market over time is gonna be much more susceptible to you know, things that are lower power, lower cost, all the things that, that aren't just performance from a silicon perspective.”
Gerstner: GCC sovereign AI compute orders will rival hyperscalers
“And I think what people still don't appreciate is they're probably dozens of sovereigns who are trying to get into the NVIDIA order book, and that they view it as one of their top three national priorities to build out AI capabilities. And I think the size you…”
Gerstner: Trillions will be spent rebuilding global compute infrastructure over 4-5 years
“But I do think that we're talking trillions of dollars over the course of the next four to five years as we rebuild the world's compute infrastructure.”
Gurley: There is no chance of competing simultaneously with TSMC and Nvidia
“And then some of them, and I, once again, I don't know that there was an exact quote, but the idea that you're going to build a fab is And compete, you're gonna compete with TSMC and NVIDIA at the same time. Like, no chance. Like, no chance.”
Gurley: Reshoring competitive semiconductor fabs to the US will not happen
“And to think you're going to re-onshore a fab and ignore Morris Chang is just kind of crazy to me. And I look at the Requirements, once again, that we put on companies around labor and say to myself, it's not going to happen here.”
Gerstner: Chip plants will be built in Vietnam, Japan, India, and Middle East
“I think you are going to see plants that get built in places like Vietnam and Japan. I think you are going to see them get built in India. You're probably going to see some attempts in the Middle East.”
Gerstner: Nvidia's $700 consensus price assumes market share drops to 26%
“The consensus forecast that has the stock at 700 dollars a share assumes, if you believe this TAM to be accurate, assumes that their share will go from 55% today to 26% in twenty-twenty-eight.”
Gerstner: AI will hit a disillusionment phase, creating a buying opportunity
“We're going to hit a zone of disillusionment where you have a mismatch between supply and demand, and then all the skeptics are going to say, I told you so. The internet's a fad. AI's a fad. Mobile's a fad. It happens in every super cycle. We're going to use t…”