Jensen Huang was delivering a speech when Donald Trump suddenly made a phone call to him, and the call was broadcast live (full text of the speech attached)

Recently, Jensen Huang, founder and CEO of NVIDIA, publicly pushed back against the rampant "AI doomsday theory" at All-In Summit 2026, labeling it as "irresponsible remarks with no scientific basis."

He listed a series of AI predictions that had been falsified over the past few years, directly pointing out that public statements from some technology companies and researchers are "creating panic and misleading the public."

Huang's target was the widely discussed blog post recently published by Dario Amodei, CEO of Anthropic — the article, co-signed by multiple cutting-edge labs, warned of AI safety risks and even quantified the probability of "civilization-level extinction."

Jensen Huang said, "This is fabricated. These people are well-educated, called researchers, and work in laboratories. The combination of these terms and this prediction is terrifying. It should not be done, and it is irresponsible."

While the summit was ongoing, U.S. President Donald Trump suddenly called the scene, directly stating that "this whole thing is a scam," and expressed firm support for data center construction, calling AI "the oil of the next 20 to 25 years."

The joint statements of the two have injected new variables into the current fierce debate around AI regulation and safety.

"All these predictions are wrong": Jensen Huang clears up doomsday predictions one by one

From the stage, Jensen Huang directly named the most widely spread AI panic predictions of the past few years, and compared them with reality one by one:

"Some people predicted that radiologists would be completely replaced by AI within five years. On the contrary, we need more radiologists than ever before." 

"Some people predicted that 90% of code would be generated by AI within six to twelve months — that was wrong. Some people predicted that 50% of entry-level jobs would disappear within six to nine months — that was also wrong." 

He also mentioned that predictions such as GPT-2 and Llama 3 once being considered "too dangerous to release" and "half of white-collar jobs will disappear next year" have all failed to come true. 

Jensen Huang's conclusion is straightforward: "These predictions are not based on science, nor are they based on research. All evidence based on science and research points to the opposite direction. This is the fear of the unknown."

He also said that for the relevant blog post recently published by Dario Amodei, CEO of Anthropic, he believes the safety-related parts in it deserve to be taken seriously, but the "scientific prediction" part about the future "obviously lacks scientific basis", "it is irresponsible for scientists to say something that is not scientific."

RSI is not "black magic": Jensen Huang explains the engineering logic of recursive self-improvement

In response to the recursive self-improvement (RSI) technology that has been hotly discussed in the market recently, Jensen Huang gave his technical judgment.

"RSI is a combination of systematic ideas," he said, "from in-context learning, skill accumulation, reflection mechanism, to reinforcement learning and synthetic data generation, these are very reasonable ideas that can enable AI to continuously make progress in problem solving."

He further explained that through low-rank adaptation methods such as LoRA, model capabilities can be improved without retraining the base model; over time, the accumulated experience will be used to retrain the base model, and the entire process is completely controllable in engineering.

Jensen Huang believes that the term "RSI" is being "weaponized" by some people to deliberately create the impression that the technology will "spiral out of control". His refutation logic is very simple: "You can do RSI inside the company, but when you release the product, you have to evaluate it, test it, and make sure there is no regression problem. The basic engineering control process is there."

Trump calls the scene: Data centers are "the oil of the next 20 years"

When the summit moved to the NVIDIA strategy breakdown session, Trump suddenly called Jensen Huang, was connected to the on-site public address system, and broadcast his speech to thousands of audiences present.

Trump said: "This whole thing is a scam. Data centers are great, they make people rich, they make states rich. It is the oil of the next 20 years, 25 years, far bigger than the Internet."

He also criticized some voices opposing data center construction for "serving people who do not want the United States to succeed", and clearly stated: "Robots will not rule the world, AI will not take over the world, the whole thing is a scam."

Trump also mentioned that about 20 trillion US dollars of investment is currently flowing into the United States, and compared this with less than 1 trillion US dollars during the previous administration.

After the call, Jensen Huang said that Trump was able to "see through this scam" and stick to his position even when polls showed opposing voices were dominant, "it takes courage."

NVIDIA's strategic logic: How far to go up, how low to press down

In terms of capital allocation and ecosystem strategy, Jensen Huang elaborated on NVIDIA's core principle: "Go as high as possible, but try to stay at the bottom layer."

He explained that NVIDIA's goal is to help the entire ecosystem succeed, rather than taking a share out of it. "If NVIDIA hadn't created cuDNN, none of the frameworks would exist. If we hadn't created Megatron Core, large-scale training wouldn't have happened. We invent the necessary technologies, then let a thousand flowers bloom."

After acquiring Hugging Face, NVIDIA's layout in the open source field has become clearer. Jensen Huang revealed that NVIDIA currently has cutting-edge open source models in five vertical fields including autonomous driving and biomedicine, including the world's first "thinking autonomous vehicle" Alpamayo, and the protein synthesis model Proteina Complexa.

He emphasized that NVIDIA's logic for entering these fields is "customers need it, and they don't have the ability to do it themselves", rather than actively disrupting competitors.

In terms of the financing ecosystem, Jensen Huang mentioned that NVIDIA is cooperating with institutions such as Blackstone and Goldman Sachs to provide financing support for AI infrastructure construction through projects such as Cloverleaf, and continuously expand the regional cloud partner network in Australia, Southeast Asia and other places.

"In some fields, superintelligence has arrived": Jensen Huang's technical judgment

At the end of the summit, Jensen Huang gave his judgment on the current stage of AI development.

The host raised the question "Have we entered the AGI moment", and Jensen Huang responded: "I think we have arrived."

He then supplemented his own definition framework: "When you look at an autonomous driving car, I don't need it to make me an omelet, I just need it to drive. In this narrow domain, it is superintelligence — the accident rate is one tenth of that of humans. For protein synthesis and virtual screening, we are already there."

Jensen Huang finally called: "I hope we can reduce the drama. Most importantly, we need to get the whole of America moving forward together. That's how we succeed."

The full text of the interview (slightly edited) is as follows:

Jensen Huang: The doomsday theory is a scam, superintelligence has arrived, the future of AI (President Trump's live connection)

All-In Podcast | All-In Summit 2026 | September 15, 2026

Host 00:00

Some people call it "foresight", which is an extremely heavy word for me, because I believe foresight is crucial. We interrupted our regular weekly program for this, and only three people can make us do that: President Trump, Jesus, and Jensen — the world's number one podcast sincerely welcomes him. He is Jensen Huang, founder, chairman and CEO of NVIDIA. Whether you realize it or not, every decision he makes is shaping your future. NVIDIA is the most important stock in this market, and Jensen can be said to be the best entrepreneur in history. The company's revenue has exploded by 97% year-on-year, and the demand is not only already strong, but actually continues to accelerate. NVIDIA is the only computing platform with a complete full-stack AI factory. The GPU is like a time machine, it allows you to see the future earlier. If we can see the future and predict the future, we have a greater chance to shape that future into its best form.

Please welcome Jensen Huang! As soon as he appeared, he won a standing ovation from the whole audience. Ladies and gentlemen, "GPU Jesus" is here, everyone loves you. Thank you, we love you too. We love the world's number one podcast, and we love your new jacket too.

Jensen Huang  01:28

I think you guys need some energy. I know we're talking about serious topics today, but we have to talk about them with full energy.

Host 01:39

Let's start with that article from this weekend. Which article?

Host Chamath  01:45

Let's start with Dario's article. This article is actually very Hemingway-style. Has anyone ever tested it with Pangram? I don't even know how much of it was written with AI assistance, but it's a pretty amazing article. What surprised many people is that the cutting-edge labs have endorsed this article and gathered around it. Jensen, can you help us sort it out — what on earth happened, and how do you interpret it? Let's dig into some of the details, and first talk about your overall opinion.

Jensen Huang 02:14

That article covers a lot of content. First of all, there is a part about safety, which we must take very seriously. Safety is the top priority, which is obvious. It is not an either-or choice between safety and leadership — you can innovate quickly, execute quickly, the United States can take the lead, and be safe at the same time. I think pitting the two against each other is a false proposition, but safety itself is of course extremely important.

The article also touches on internal control issues, which I think is what Dario is talking about. Obviously, what the whistleblower named Coxson exposed is very serious — as long as someone reports it, it must be taken seriously. I think Coxson had great courage to make his concerns public. Even so, some of the issues in those concerns are mixed together.

The act of whistleblowing itself, I think, is right. But the scientific predictions about the future in the article are not so tenable, because they are not based on a scientific foundation. The author of this article is a scientist, but the content obviously has no scientific basis, which I disagree with.

As for the suspension and slowdown of development mentioned in the article, these are things they can do voluntarily — if they feel the company is out of control. If Coxson sees signs of loss of control — of course we don't know what he saw specifically — maybe it's related to the transition from research to engineering. As we all know, these labs are transforming from research institutions to engineering organizations, with extraordinary talents and excellent engineering capabilities, but engineering is different from research, and this transformation process will inevitably have some stumbles. We don't know what he saw exactly, and only he knows in the end.

But if the problem really lies in the lack of control, that's another issue. How should the government respond? The word "regulation" covers all the problems at once, and everything is in a blog post.

Host  Chamath 04:35

Can you help us sort it out? For example, the term "civilization extinction" — I don't even know how to explain it to my mom. When very smart people quantify these concepts, such as "10% probability of extinction", that's probably what makes many people uneasy. No one knows how to explain to ordinary people what this means, or how this could happen.

Jensen Huang 04:56

We shouldn't say that, because it's completely made up. These people are well-educated, they are called researchers, and they work in laboratories. But the combination of these terms and this prediction is shocking, irresponsible, and should not be done.

Jensen Huang 05:23

Let's go back to the real facts. Some people once predicted that radiology would be completely replaced by artificial intelligence within five years, and there would be no radiologists in the world. It turns out that the opposite is true — we need more radiologists than ever. However, AI has indeed fully penetrated the field of radiology, it has automated scan reading, which is great.

Others predicted just last year that 90% of code would be generated by AI within six to twelve months — which turned out to be wrong. Last year, others predicted that 50% of entry-level jobs would disappear within six to nine months — also proved wrong. Saying GPT-2 is too dangerous to release, saying Llama 3 is too dangerous to release... We've heard that half of white-collar jobs will disappear next year, that there will be an "employment doomsday". All these predictions have been proven wrong. We must be accountable for these absurd predictions, someone must take responsibility for them, and we should record all of this. Of course, people are indeed recording, and they will remind us — these predictions are contradictory to the United States winning the AI competition.

Host Chamath 06:56

Simply put, some people will say "trust the experts" and cite the COVID-19 pandemic as an analogy — it all started with researchers, educated people who had asymmetric knowledge about this matter that no one else had, and then said something that was ultimately proven incorrect by facts. So now there is a war between the "trust the experts" movement and the camp that says "let's look at the actual historical record of these predictions, let's think more systematically".

What do you think is behind this? Is it commercial interests or political interests? Why do you think they did this?

Jensen Huang 07:44

First of all, I would say these are some of the most influential companies in history. They have top engineers, outstanding researchers, and have done truly remarkable work. On the one hand, I have a very close partnership with them as a company; on the other hand, it is really a pity that we have to have such a conversation in public. I think these companies should operate the way we used to build companies — quietly.

Speaker 3 08:11

Jensen, you don't allow anyone in the company to speak on behalf of the entire organization, especially when they are in a bad state or acting on impulse, they can't represent you and the company to tweet casually, right?

Jensen Huang 08:31

That's right, because when they join our company, we tell them — this is the code of conduct for you to work at our company. If you identify with NVIDIA's culture — everyone knows that NVIDIA's culture and employee satisfaction are extremely high — employees love the stability of the company, love that our core values are always consistent: take care of employees' families, create conditions for them to complete their life's work, do meaningful work, try to keep a low profile, and contribute to everyone's success. These core values will attract people who identify with them to join.

But when you come to work at our company, there are also some things we don't welcome. For example, we don't welcome political discussions inside the company. On topics such as race, religion, and politics, we tell employees to leave those talks outside the company. NVIDIA is a de-politicized company, we are cross-partisan, and we want the United States to succeed. No matter which administration is in power, we will do everything we can to help the United States succeed.

Chamath  09:51

Then on the issue of AI regulation, more specifically, Satya is here this morning, and his statement is: before discussing regulations that may greatly hinder development, why not get the basics right first? Why not get the assessment right and get standardization right