“Dynamo, as you know, is a piece of instrument, a machine that was created by Siemens to turn essentially water into electricity. And Dynamo uh powered the…”
Physical-AI and AI-factory industrialization plays straight into Siemens' automation franchise
“the alternatives, the custom ASIC, AMD, others are going to cost 25 to 30 billion, and you're going to lose share.”
Even free chips lose if they can't match Nvidia's pace, a direct knock on AMD's cheaper alternative
“You announced one here. Uh, I believe it's the Dell 6800. Uh, this is a very powerful workstation to run local models, 750 gigs of RAM.”
Local-model mania drives demand for Nvidia-powered Dell AI workstations to run open-source agents
“They're going to be agents banging on SQL, they're going to be agents banging on vector databases, agents banging on Blender, agents banging on Photoshop.”
Agents banging on Photoshop reframes Adobe as a consumption beneficiary, not an AI casualty
“in the case of chip design, Synopsys and Cadence and uh Omniverse and Blender and Autodesk and all of these tools are going to continue to be used.”
Agents bang on EDA tools 100x harder, busting the seat-count ceiling for Synopsys and Cadence
“Omniverse and Blender and Autodesk and all of these tools are going to continue to be used.”
Autodesk's design tools persist as agent control surfaces rather than getting displaced
“I want everything to be put back into Cadence, because that's how I control it, that's how I ground truth.”
Cadence is the control-layer chip-design tool agents will pound at far higher volume
“there was a training run that happened in this crypto project called Bit Tenser. Subnet 3, they managed to train a 4 billion parameter Lama model totally…”
Brad points to Bittensor's subnet-3 distributed training run as proof that decentralized training is real, which is the core bull thesis for TAO, the network's token. The directional read is long: validating the technology raises the network's credibility and demand for TAO.
“with Uber that you're going to have a number of cars on the road for many different manufacturers.”
Nvidia's open AV stack feeds an Uber fleet across many manufacturers, an asset-light AV aggregator
“There was just a video of you driving around in a Mercedes.”
Mercedes is a flagship Nvidia Drive partner shown running the reasoning AV system
“You've added many more partners including BYD.”
BYD joins Nvidia's self-driving partner roster, riding China's robotics/EV supply-chain dominance
“In the case of Elon and Tesla, they buy our training computers.”
Humanoid robots hit real products in 3-5 years, with Optimus a named frontrunner on Nvidia training compute
“Google has TPU, Amazon has Inferentia and Trainium. You know, everybody's sort of spinning up their own version of I think I can out Nvidia Nvidia, even…”
Google's TPU is the most credible in-house challenger to Nvidia and a Gemini cost lever
“In the case of AWS, I think they just announced, I think it was yesterday, that they're going to buy a a million chips uh in the next couple of years. I mean…”
AWS just committed to a million Nvidia chips on top of its own Trainium/Inferentia buildout
“Yeah, first of all, they just don't understand the scale and the breadth of AI... Most people think that AI is in the top five hyperscalers.”
Jensen rejects the consensus model that has NVIDIA growth decelerating to 30%, then 20%, then 7% by 2029, arguing analysts cannot see that AI extends far beyond the top five hyperscalers. The buy thesis is that street estimates are anchored to law-of-large-numbers orthodoxy that NVIDIA's actual order book will break.
“open AI is incredible. They're going to be enormous. Anthropic is incredible. They're going to be enormous. But AI is going to be much, much bigger than that.”
Jensen, whose chips power the entire frontier-model buildout, singles out OpenAI as the category leader headed for enormous scale. The cleanest way to express that named view on-venue is the OpenAI private-valuation perp, which prices the implied company valuation rather than a public equity.
“Dario on Dkesh's podcast recently said by 2027, 28, we'll have hundreds of billions of dollars of revenue out of the model companies and the agent companies,…”
Jensen says Dario's trillion-by-2030 is conservative; every enterprise app becomes an Anthropic reseller