Jensen Huang

Huang predicts every European company will become an AI company, producing a thriving AI economy.

Everybody can build agents. Every company in Europe will become an AI company … and we’re going to have a thriving AI economy here in Europe.

Huang asserted that AGI has already arrived, citing the launch of OpenAI's GPT-6 Astra built on Nvidia GPUs.

Huang reiterated that assessment in September, saying that "AGI had arrived" with the launch of Open-AI's GPT-6 Astra, which was built on more than 100,000 Nvidia Grace Blackwell GPUs.

Huang predicts AI will enable understanding of the biological machine within about five years, making the end of disease a reasonable expectation.

Understanding the biological machine is right around the corner. It's, it's not 10 years. It's five years probably.

Responding to 'AI slop' criticism, Huang says DLSS 5 is guided by artist-defined geometry and textures and is an optional tool for developers.

But DLSS 5 is 3D conditioned, 3D guided. It's ground truth structure data guided. And so the artist determined the geometry.

Huang says intelligence is being commoditized and humanity, character and compassion are the true superpowers.

I actually think intelligence is a commodity. I'm surrounded by intelligent people. And I'm surrounded by intelligent people more intelligent than I am in each one of the spaces that they're in.

Huang advises every worker and student to become expert in AI, warning those whose job is merely a task will be disrupted.

If we were to hire a new college graduate today, and I have a choice between two, one that have, that is no clue what AI is and one that is expert in using AI, I would hire the one who's expert in using AI.

Huang argues coding is now specification, expanding the number of coders from 30 million to roughly 1 billion.

Describe a specification for a computer to go... Telling the computer what to go build. How many people? I think we just went from 30 million to probably 1 billion.

Citing radiologists' growth despite superhuman computer vision, Huang says NVIDIA's software engineer headcount will grow.

The number of software engineers at NVIDIA is gonna grow, not decline. And the reason for that is because the purpose of a software engineer and the task of a software engineer coding are related, not the same.

Under a definition of AI building a billion-dollar company, Huang says AGI is achieved now, though agents building an NVIDIA is zero percent likely.

I think it's now. I think we've achieved AGI.

Huang says NVIDIA could reach $3 trillion in revenue in the near future, calling its growth inevitable.

is it possible for NVIDIA to be a, you know, $3 trillion revenues company in the near future? The answer is of course yes. And the reason for that is because it's not limited by any physical limits.

Huang is certain the share of GDP spent on computation will rise 100-fold as computers become revenue-generating factories.

I'm absolutely certain the percentage of that GDP that will be used for computation will be 100 times more than the past... because it's no longer a storage unit. It's a product generation unit.

Huang predicts tokens will segment into tiers like iPhones, with premium tokens soon costing $1,000 per million.

You know, the idea that somebody's willing to pay $1000 per million tokens is just around the corner. It's not if, it's only when.

Huang names CUDA's install base, amplified by annual execution velocity and a broad ecosystem, as NVIDIA's biggest moat.

Our single most important property as a company is the install base of our computing platform. Our single most important thing today is our, is the install base of CUDA.

Huang argues TSMC's edge lies beyond technology in its orchestration of dynamic demand, customer-service culture and trust.

You know, first of all, the deepest misunderstanding about TSMC is that their technology is all they have. That somehow they have a really great transistor, and if somebody shows up another transistor, game over.

Huang says NVIDIA will keep building open models like Nemotron to diffuse AI across industries, including biology and physical AI.

And if everything is proprietary, it's hard to do research and it's hard to innovate on top of, around, with. And so...Open source is fundamentally necessary for many industries to join the AI revolution.

Huang calls China the fastest-innovating country, citing talent, open source culture and fierce inter-provincial competition.

So you get this rapid, incredible great talent, rapid innovation because of open source and just, you know, the nature of friends, and insane competition. Among comp- among the company, what emerges is incredible stuff. And so this is the fastest innovating country in the world today

Huang notes about half of the world's AI researchers are Chinese, mostly still in China.

50% of the world's AI researchers are Chinese, plus or minus, and they're mostly in China still.

Huang proposes tapping idle grid capacity by having data centers accept flexible power contracts and gracefully degrade during peak demand.

And so 99% of the time, our power grid has excess power, and they're just sitting idle, but they have to be there sitting idle because just in case, when the time comes, hospitals have to be powered

Asked about ASML EUV and TSMC CoWoS bottlenecks, Huang says he is not worried because suppliers have committed to meet NVIDIA's needs.

No. - Okay. - Because, because I told 'em what I needed. They understood what I need. They told me what they're gonna go do, and I believe them what they're going to do.

Huang says he persuaded DRAM CEOs to invest in HBM and LPDDR for data centers, and all three memory makers had record years.

About three years ago, I was able to convince several of the CEOs that even though at the time HBM memory was used quite scarcely, you know, and, and barely by supercomputers that this was going to be a mainstream memory for data centers in the future.

Huang says NVIDIA scaled computing a million times in 10 years versus Moore's Law's 100x, driving token costs down an order of magnitude annually.

You know, the our computer price is going up, but our token generation effectiveness is going up so much faster that token cost is coming down. It's just coming down an order of magnitude every year.

Huang dismisses the idea that AI will destroy software, arguing agents will use existing tools like a humanoid robot would use a microwave.

You know, a lot of people would say, "You know AI is gonna completely destroy software. We don't need software anymore. We don't even need tools anymore." That's ridiculous.

Huang says OpenClaw is the ChatGPT moment for agentic AI, calling it the 'iPhone of tokens'.

And I think OpenClaw did for, did for agentic systems what ChatGPT did for generative systems. And I just think it's a very big deal.

Huang outlines a fourth 'agentic' scaling law in which agents spawn sub-agents, and says intelligence will ultimately scale with compute.

And so the next scaling law is the agentic scaling law. It's kind of like multiplying AI. Multiplying AI, we could spin off agents as fast as you want to spin off agents.

Huang argues inference is not a cheap, commoditizable workload because test-time reasoning is intensely compute-intensive.

And, and that was always illogical to me because inference is thinking, and I think thinking is hard. Thinking is way harder than reading.

Huang rejects the 'out of data' concern, arguing synthetic data means training is now constrained by compute rather than data.

Training is no longer limited by... Data is now limited by compute. And the reason for that is most of the data is synthetic.

Huang explains that extreme co-design across CPU, GPU, networking and switching is necessary because AI workloads no longer fit in one computer and Moore's Law has slowed.

And so distributed computing at the scale that we do, the CPU is a problem, the GPU is a problem, the networking is a problem, the switching is a problem. And distributing the workload across all these computers is a problem. It's just a massively complex computer science problem. And so we just gotta bring every technology to bear.

Huang rejected Geoffrey Hinton's 10%+ catastrophic AI risk estimate as unscientific, saying such predictions lack research support.

"Just because it comes from a scientist doesn't make it scientific," Huang said, arguing that such predictions were not supported by research.

Huang argued that Germany's manufacturing and engineering expertise is irreplaceable and that AI will augment rather than replace it.

That knowledge belongs to you,” Huang said. “It cannot be replaced. AI can only augment your capability, it will not replace your capability. Take advantage of the technology. It’s a new instrument.

Huang described AI as a mighty new industrial revolution and urged Germany to seize the chance to re-engineer its economy.

This industrial revolution is so mighty, so powerful,” Huang said. “In the Chancellor’s words, you are re-engineering Germany. This is your chance to do it well.

Jensen Huang warned that Germany is falling behind in the AI race because excessive hesitation and analysis are holding Europe back, urging policymakers to deploy AI faster.

Despite its strong potential and wealth of talent, Huang believes progress in Europe is being blocked by too much hesitation and analysis and has called on policymakers to be quicker to put AI into practice.

Jensen Huang says current AI lacks physical reasoning, and physical AI embodied in robots should make the next generation of factories highly robotic within ten years.

Most AI today doesn’t understand the laws of physics, friction, inertia, or cause and effect. These are physical reasoning abilities that children and even our pets have… When you put physical AI into a physical object, you get robotics. Hopefully, in the next 10 years, this new generation of factories will be highly robotic.

Huang says neocloud demand is off the charts and the capital shortage constraining growth will be solved.

What is missing in the middle is the working capital to fuel that growth… because there's just so much demand across so many sectors of AI, there's a scarcity of capital. The demand of the capital is much greater than the availability of the capital. That's going to get solved.

Huang told German Chancellor Merz that regulation is meaningless without a thriving industry and that Germany must build a domestic AI sector safely.

"There's nothing to regulate if the industry doesn't thrive," Huang told Merz, arguing that Germany needed to foster a domestic AI industry while ensuring it developed safely and responsibly.

Huang argued policymakers should regulate AI applications, as with internet-based products, rather than the underlying technology wholesale.

"We should be regulating the applications right away," Huang said.

Huang said the next AI wave is physical AI in robotics and industry, and Germany's manufacturing strength can be an advantage.

"The next wave of AI is physical," Huang said, referring to robotics, manufacturing, cars, and heavy industry. "Electrons will never replace atoms."

Huang told German businesses that AI will strengthen rather than replace the country's industrial capabilities, describing it as a new instrument for companies.

AI can only augment your capability, it will not replace your capability

Huang argues AI will augment rather than replace human and industrial capability, urging Germany to adopt it.

AI can only augment your capability. It will not replace your capability. So take advantage of the technology

Huang urges Germany to use its manufacturing data and expertise to train AI kept within Germany, framing it as sovereign AI that can re-engineer the economy.

take advantage of artificial intelligence. Let them codify, let them program, let them teach this artificial intelligence and keep it here in Germany, and let’s re-engineer, the chancellor’s words, Germany. This is your chance to do it.

Huang says the next wave of AI is physical AI, spanning robotics, manufacturing, cars and heavy industry.

The next wave of AI is physical. This is where AI interacts with the physical world: robotics, manufacturing, cars, heavy industry.

Huang argues AI will augment rather than replace Germany's manufacturing capabilities, urging adoption.

And so AI can only augment your capability. It will not replace your capability. So take advantage of the technology.

Huang warns Germany that the AI industrial revolution is a one-time chance it cannot afford to miss.

This industrial revolution is so mighty, so powerful. You don't have another chance,

Huang advises Germany to import AI technology while keeping its culture, values and knowledge at home, addressing data sovereignty concerns.

you should not export your culture, export your values, export your knowledge.

Huang describes Germany's manufacturing data and expertise as a national natural resource for AI.

is your natural resource, it belongs to the nation

Jensen Huang says useful quantum computing requires highly complex fault-tolerant systems and that GPU supercomputers working with quantum processors can accelerate discovery of transformative quantum applications.

Useful quantum computing will require fantastically complex, fault-tolerant systems — and understanding how they behave at scale is essential to building them. Our results from Jülich show how quantum processors and GPU supercomputers can work together to tackle research that would otherwise be out of reach — accelerating the discovery of transformative quantum applications.

Huang told Bloomberg that recent wobbles in AI stocks will pass because AI will positively impact every industry and every country.

going to impact in a very positive way every single industry and every single country

Jensen Huang warned that demand for capital far exceeds its availability amid AI investment, but said the gap will be resolved.

“The demand for capital is much greater than the availability of capital,” Jensen said, adding: “That’s going to get solved.”

Jensen Huang says RTX Spark, built with Microsoft and PC partners, ushers in a new generation of Windows PCs designed for personal AI agents as well as today's creation and gaming work.

Together with Microsoft and our PC industry partners, we’re bringing creators, developers and gamers a new generation of Windows PC, built for the work they do today and the personal AI agents they’ll work with next.

Huang argues reasoning-based AI will require a billion times more compute than before because thinking and problem solving need far more inference.

First of all it's very fairly logical why the amount of compute that we're going to need uh is a billion times more than where we were just recently so calla last year. And the reason for that is because thinking and problem solving takes a lot more you know uh inferences

Huang says memory management and tool use in model harnesses were the biggest unlocks for AI agents over the past year.

so these two areas knowledge knowledge um and working memory and um tool use huge unlocks in the last year.