Brad Gerstner

1. Key Themes

The AI Factory Paradigm Shift: From GPU Company to Full-Stack Infrastructure

NVIDIA has fundamentally redefined itself. Jensen describes an evolution from a single-rack GPU company to a multi-rack AI factory company spanning GPUs, CPUs, networking, storage, and now LPU processors via Groq. This isn't just product expansion — it's a TAM expansion of 33–50%.

"We used to be a one rack company. We now added four more racks. So NVIDIA's TAM if you will, increased from whatever it was to probably something, call it, you know, 33%, 50% higher." 00:04:22

"We just really evolved from a GPU company to an AI factory company." 00:02:49

The critical insight for investors: the cost of the factory is not the cost of the tokens. Jensen explicitly argues the most expensive factory produces the cheapest tokens due to throughput efficiency.

"The big idea is that you should not equate the price of the factory and the price of the tokens, the cost of the tokens. It is very likely that the 50 billion dollar factory, and in fact, I can prove it, that the 50 billion dollar factory will generate for you the lowest cost tokens." 00:07:33

"Even when the chips are free, it's not cheap enough." 00:08:49

The Agentic Computing Revolution: A 10,000x Compute Expansion in Two Years

Jensen frames the progression from generative AI → reasoning → agentic as two successive 100x jumps in compute demand, totaling a 10,000x increase in just two years. Crucially, he believes this has barely begun, with a millionx being the ultimate destination.

"When we went from generative to reasoning, the amount of computation we needed was about 100 times. When we went from reasoning to agentic, the computation is probably another 100 times. Now we're looking at, in just two years, computation went up by a factor of 10,000x." 00:22:26

"We haven't even started scaling yet. We are absolutely at a millionx." 00:23:20

OpenClaw is identified as the cultural inflection point — the moment agentic AI entered public consciousness, analogous to what ChatGPT did for generative AI.

"OpenClaw basically put into the popular consciousness what an AI agent can do. That's the reason why OpenClaw is so important, from a cultural perspective." 00:14:27

Physical AI: A $50 Trillion Industry Being Touched by Technology for the First Time

Jensen identifies Physical AI — robots, autonomous vehicles, and intelligent physical systems — as the technology industry's first real opportunity to penetrate a $50 trillion sector that has largely been untouched by software. This is a 10-year journey now inflecting.

"Physical AI as a large category. It's technology industry's first opportunity to address a $50 trillion industry that has largely been void of technology until now." 00:10:46

"It is a multi-billion dollar business for us. It's close to $10 billion a year now. And so it's a big business and it's growing exponentially." 00:11:14

Robots are expected to be commercially prevalent within 3–5 years, driven in part by China's hardware superiority in motors, magnets, and rare earth materials — a supply chain dependency that is under-discussed.

"From the point of high-functioning existence proof to reasonable products, technology never takes more than a couple, two, three cycles... somewhere around three years to five years." 00:52:57

2. Contrarian Perspectives

Enterprise Software Won't Be Destroyed by AI — It Will Be Amplified 100x

The prevailing narrative is that AI agents will gut enterprise software companies. Jensen inverts this: agents will become the most prolific consumers of enterprise software tools, driving a 100x increase in seats/usage — not replacement.

"The enterprise software industry is limited by butts and seats. It's about to get 100 times more agents banging on those tools. There are going to be agents banging on SQL. There are going to be agents banging on vector databases, agents banging on Blender, agents banging on Photoshop." 00:30:08

"Those tools are the conduit between us. In the final analysis, when the work is done, it has to be represented back to me in a way that I can control." 00:30:33

Dario Arevalo's Trillion Dollar AI Revenue Forecast is Too Conservative

Brad Gerstner cited Dario's forecast of hundreds of billions in model/agent revenue by 2027-28, reaching $1 trillion by 2030. Jensen called this dramatically understated.

"I think he's being very conservative. I believe Dario and Anthropic is going to do way better than that." 00:56:35

His reasoning: every enterprise software company will become a value-added reseller of AI model tokens, creating a distribution flywheel that the model companies themselves haven't fully priced into forecasts.

"I believe every single enterprise software company will also be a reseller, value-added reseller of Anthropic's tokens... their go-to-market is going to expand tremendously this year." 00:57:06

AI Doomerism Is America's Biggest National Security Threat — Not AI Itself

Jensen argues that the greatest risk to U.S. national security from AI is not misuse of AI, but rather the U.S. failing to adopt it out of fear — mirroring what happened to the nuclear and solar industries.

"Our greatest source of national security concern with respect to AI is that other countries adopt this technology while we are so angry at it or afraid of it, or somehow paranoid of it, that our industries, our society don't take advantage of AI." 00:17:54

Brad reinforces this with a stark historical analog: the U.S. effectively shut down nuclear power, and now China has 100 fission reactors being built versus zero in the U.S.

"17% popularity of AI in the United States. I mean, we see what happened to nuclear, right? We basically shut down the entire nuclear industry. And now we have a hundred fission reactors being built in China and zero in the United States." 00:20:24

Technology Adoption Eliminates Tasks, Not Jobs — The Radiologist Example

Contrary to the mainstream narrative that AI will eliminate professional roles, Jensen provides an empirically grounded counter-example: a world-class computer scientist predicted radiology would be eliminated by computer vision. It wasn't. The number of radiologists actually increased because faster scanning enabled more volume, more revenue, and earlier detection.

"Ten years later, his prediction was at 100% right. Computer vision has been integrated into all of the radiology technologies... The surprising outcome is the number of radiologists actually went up and the demand for radiologists has skyrocketed." 00:03:26

CUDA's Ubiquity Across All Clouds and Edges Is an Underappreciated, Near-Unassailable Moat

Most analysts focus on chip-level comparisons. Jensen identifies a structural advantage that transcends chip pricing: CUDA is the only architecture deployable across every cloud, on-prem, in cars, at the edge, and in space — simultaneously.

"We're the only architecture that could be in every cloud. And that gives us some fundamental advantages. We're the only architecture you could take from a cloud and put into on-prem, in the car, in any region. In space." 00:43:27

"40% of our business... the customers don't know what to do with you. They're not trying to build chips. They're not trying to buy chips. They're trying to build AI infrastructure." 00:43:27

3. Companies Identified

NVIDIA

Full-stack AI infrastructure company spanning chips, networking, simulation, robotics, and software platforms. Central to the entire conversation. Jensen argues NVIDIA is actually gaining share despite custom ASIC competition, with AWS reportedly purchasing 1 million chips in coming years atop existing orders.

"We're the only AI company in the world that works with every AI company in the world. They never show me what they're building. And I always show them exactly what I'm building." 00:42:57

Anthropic

AI safety-focused frontier model company. Jensen is an explicit fan of their technical excellence and safety culture. Believes their revenue trajectory will dramatically exceed Dario's own forecast.

"The technology is incredible. We are a large consumer of Anthropic technology. Really admire their focus on security, really admires their focus on safety." 00:18:51

Groq

LPU (Language Processing Unit) maker acquired by NVIDIA. Jensen recommends allocating Groq LPUs to ~25% of Vera Rubin data center space for disaggregated inference workloads — specifically the decode-heavy portions of inference pipelines.

"We should add Groq to about 25% of the Vera Rubins in the data center." 00:03:16

Open Evidence

AI company in healthcare diagnostics. Cited by Jensen as a strong example of agentic AI transforming how patients interact with medical systems.

"Open Evidence is a really good example." 00:50:25

Hippocratic AI

Healthcare AI agent company. Cited alongside Open Evidence as leading the transformation of healthcare interactions.

"Hippocratics is a really good example. Love working with those companies." 00:50:25

Ohalo (David Friedberg's company)

Genomics/agriculture company. Demonstrated a PhD-level scientific discovery in 30 minutes using auto-research tools — a result that would normally take 7 years and be published in a top journal.

"Something was published internally that we said, oh my God. And that would normally be a PhD thesis that would take seven years... And it was done in 30 minutes on a desktop computer running on auto research." 00:28:11

BitTensor / Subnet 3

Decentralized crypto-AI project that successfully completed a stateful distributed training run on a 4 billion parameter LLaMA model using contributed excess compute. Jensen compared it to the modern version of Folding@home.

"Our modern version of folding at home." 00:31:29

4. People Identified

Peter Steinberger

Creator of OpenClaw (open-source agentic computing framework). Jensen describes OpenClaw as not merely a tool but as a blueprint for the operating system of modern personal AI computing. NVIDIA is actively contributing engineering resources to help secure it.

"Peter Steinberger was here. And so we've got a mound of great engineers working with him to help secure and keep that thing so that it could protect our privacy, protect our security." 00:16:37

Dario Amodei (Anthropic CEO)

Referenced for his trillion-dollar AI revenue forecast by 2030, which Jensen believes dramatically understates the real trajectory.

"I think he's being very conservative. I believe Dario and Anthropic is going to do way better than that." 00:56:35

Altimeter Capital founder, praised by Jensen for his work with policymakers on AI education and his framing of the AI diffusion challenge.

"Brad, you do a great job doing this. We had to get in front of them and inform them about the state of the technology, what it is, what it is not." 00:16:59

5. Operating Insights

Engineers Who Don't Consume AI Tokens Proportionate to Their Salary Are Underperforming

Jensen articulates a concrete framework: if you pay an engineer $500K/year, they should be consuming at minimum $250K in AI tokens annually. Spending only $5K in tokens is equivalent to a chip designer refusing to use CAD tools.

"If that $500,000 engineer did not consume at least $250,000 worth of tokens, I am going to be deeply alarmed. This is no different than one of our chip designers who says, I'm just going to use paper and pencil. I don't think I'm going to need any CAD tools." 00:24:51

Operators should build internal KPIs around token consumption per employee as a leading indicator of AI adoption effectiveness.

Deep Vertical Specialization Is the Only Durable Moat at the Application Layer

In a world where general models are commoditizing rapidly, Jensen's explicit advice is that application-layer moats come from vertical depth and proprietary data flywheels — not horizontal generalization.

"Deep specialization... The sooner you connect your agent with customers, that flywheel is going to cause your agent to get hyper-specialized." 00:57:37

"Know your vertical. Know it as deep and as better than everybody else. And then wait for these tools because they're catching up to you, and now you can imbue it with your knowledge." 00:58:03

The CEO's Primary Job Is to Find the Intersection of "Insanely Hard" and "Unique Company Superpowers"

Jensen's strategy filter is explicit and actionable: only pursue initiatives that (1) have never been done, (2) are insanely hard, and (3) require the unique capabilities of your specific organization. If something is easy, withdraw — competition will follow.

"Is this something that has never been done before that's insanely hard to do? And that somehow taps into the special superpowers of our company. And so I have to find this confluence of things that meets the standard." 00:09:54

6. Overlooked Insights

Telecommunications Base Stations Are About to Become AI Edge Infrastructure — A $2 Trillion Industry Transformation Nobody Is Discussing

Jensen very briefly mentioned that NVIDIA is working to turn telecom base stations into AI infrastructure nodes. This received almost no follow-up, yet it represents a $2 trillion industry being quietly repurposed as AI edge compute. Every radio tower becoming an AI edge device would represent one of the largest infrastructure repurposings in history — and NVIDIA is already working on it.

"One of the most important ones is one that we're working on that basically turns the telecommunications base stations into part of the AI infrastructure. So now all of the, it's a $2 trillion industry. All of that in time will be transformed into an extension of the AI infrastructure. And so radios will become edge devices, factories, warehouses, you name it." 00:06:14

This is a massive signal for investors watching the intersection of telecom, edge compute, and AI infrastructure — yet it passed in roughly 15 seconds of conversation.

Digital Biology Is at Its ChatGPT Moment — Jensen Believes Healthcare Will Inflect Within 2–5 Years, Not 10–20

Jensen very matter-of-factly stated that digital biology is near its ChatGPT-equivalent inflection point — the moment when the underlying capability becomes universally accessible and adoption explodes. Most biotech timelines assume a decade-plus horizon. Jensen is calling it in 2–5 years.

"I think we are literally near the ChatGPT moment of digital biology. We're about to understand how to represent genes, proteins, cells... In five years time, I completely believe that the healthcare industry, where digital biology is going to inflect." 00:11:14

Given that Ohalo demonstrated a 7-year PhD thesis in 30 minutes using auto-research in genomics on the same podcast, this is not hypothetical — there is live evidence of the inflection already occurring in private.