Jensen Huang's Bloomberg Interview with Emily Chang: Summary and Timestamps

A summary of The Rise and Reckoning of Jensen Huang, Emily Chang's Bloomberg film: AI risk, China, containing AI agents, circular financing and who runs Nvidia next.

▶ Watch the full interview on YouTubeBloomberg Originals · 12 November 2026

The Rise and Reckoning of Jensen Huang is an episode of The Circuit with Emily Chang, published by Bloomberg Originals on YouTube on 12 November 2026^[1]. It runs 45 minutes and had more than 3.3 million views by 13 November 2026. Chang filmed it in October 2026, when she followed Jensen Huang, then Nvidia's CEO, to four countries in four days: France, the UK, Sweden and Germany. Weeks later, reports about his relationships made him the focus of internal and government investigations. The video came out the day Nvidia announced that he had resigned and that its chief financial officer, Colette Kress, was interim CEO^[2].

This page summarises what Huang said, by topic, with timestamps, and marks what comes from Bloomberg's narration.

Topics and timestamps

The chapter names are taken from the episode description^[1].

The Europe tour and Nvidia's investments

The tour started in Paris. Huang said Nvidia had invested $8 billion in European startups in the past year and would announce another two billion dollars the following week. Bloomberg's narration describes his idea of AI as a five-layer cake: energy, chips, infrastructure, models and apps. It says Nvidia started in the chip layer and now invests in every layer, in power, data centres, AI labs and startups. The narration also says Nvidia agreed to buy Hugging Face for $12.9 billion.

In London, Huang said Nvidia had put him in the same hotel 27 years earlier, when the company went public. He said he still does not know how to write a business plan. Asked what would make the trip a success, he said he has no end game, roadmap or plan, and that for him "the journey is the destination".

The narration then gives his background. He was born in Taiwan in 1963, moved to the US, washed dishes at Denny's and started at Oregon State University at 16. He designed chips at AMD and co-founded Nvidia in 1993 to make graphics chips for video games. In 2006 Nvidia released CUDA, software that let its chips do work beyond graphics. AI researchers found the chips could train their models, and Huang went all in on AI before most of the industry caught on.

Is there zero chance AI ends humanity?

At the Paris event, Huang was asked whether he was saying there is zero chance AI ends humanity, because it is only an engineering problem. He said no: what he says is that it is in our control. He said advancing capability and advancing safety are not in conflict but are one thing.

At Nvidia's GTC conference in Berlin, Chang said he had put the chance of AI ending the world by 2030 at 0%, and asked about 2040. Huang said he was "quite sure", because he is confident in the agency of people. He said the world today is safer than it was 50, 100 or 250 years ago, and that the facts show it. He said news headlines are more pessimistic because bad news is easier to sell.

Trump, superintelligence and self-regulation

Chang asked what happens when his views differ from President Trump's. Huang said they do not have to agree on anything. He said his job is to inform the president about how decisions affect Americans and the world, whether or not the decision serves his own interests.

The narration says that in late September Huang joined a White House summit where Trump called AI "super intelligence". The White House presented an accord that puts the tech industry in charge of voluntarily policing itself. Chang noted that Huang had spent three months saying AI is not a new species but code. Huang said "super" is better than "artificial", and that Trump's point was that "artificial" just says fake. He said AI does a better job of researching and writing than he does, but is not better at everything people do.

The film quotes a critic who called the pledge "great for the cameras, but utterly meaningless". Chang asked whether builders should help write their own rules. Huang said this is "very appropriate". The industry should hold itself to the highest standards before it expects governments to regulate it. He compared the plan to the internal and external safety audits used in many industries. Even with government regulation, he said, someone needs the ability to do the auditing, so that industry and its startups have to be created.

Sovereign AI on American chips

The narration says Huang is making the term sovereign AI popular: the idea that every country should build and control its own AI. It notes that the more sovereign AI there is, the more demand there is for Nvidia chips. Chang asked whether AI is still sovereign if it runs on American chips. Huang said yes. Sovereignty means a country controls its own data, culture, values and institutional knowledge, and decides its own future. A country can start with someone else's technology or open models, then adapt and fine-tune them.

In London he said an autonomous system asked to make an omelette does not need access to a chainsaw or a shotgun. He praised Nscale, a London data centre company that Nvidia backs, as possibly the most valuable UK startup of the last ten years. According to the narration, a few days later investors questioned Nscale's valuation, and Microsoft pulled out of a deal with the company over the cost of electricity in the UK. Huang said the UK was at a "Goldilocks moment", but the country tends to overthink: given a million ways to win, it finds the one reason not to act.

Asked whether the world needs something like a UN for AI, Huang said "maybe", and that dialogue has to come first. He said every country wants to be part of the dialogue about intelligence.

China and the Alice Huang reports

The narration says Huang has argued that limiting China's access to US technology will make China's own chip industry stronger. Chang asked whether the Chinese market is slipping away. Huang said it is "completely blocked from us today". He said he did his best to tell US policymakers why Nvidia should serve that market for as long as it can, and that he thinks the US made a mistake that is hard to unwind. He said there are more Chinese AI researchers than researchers from any other nation. China competes with the US but is not its enemy, he said, and if the US cares about AI safety it must open dialogues with China.

The next segment was added after the interview was recorded. The narration says US intelligence has reportedly identified Alice Huang, also known as Huang Lu, a former television journalist, as a spy recruited by Beijing to target the Nvidia CEO. She is an executive at Megaspeed International, one of the largest buyers of Nvidia chips in Southeast Asia, which US officials suspect of smuggling advanced chips to China. Huang first called their interactions normal, and later admitted an affair in a memo to employees. The film quotes the memo: "I was unfaithful to my wife." The narration also says Bloomberg reported that in 2025 Nvidia intervened in a US government probe of suspected chip smuggling by Megaspeed. The film notes that Nvidia's board says it has no reason to believe company secrets or US national security were compromised, and that China calls the allegations a smear.

Back in the October interview, Chang asked what China could do that would make him stop selling chips there. Huang said the Chinese are pragmatic people who want their country to be rich, and the US should respect that aspiration. He said the nations to worry about are the ones that do not care about their people's prosperity.

Containing and monitoring AI agents

The narration says OpenAI had disclosed six new incidents of what it calls model misalignment, and Anthropic said a Claude model had carried out unintended actions. Nvidia released OpenShell, an open-source platform to govern AI agents and shut them down if needed.

Chang said Sam Altman, Dario Amodei and Elon Musk all say AI should slow down, and asked what Huang sees that they do not. He said capability and safety advance together. His example was hallucination in the first ChatGPT. Making chatbots grounded in truth needed new technology: retrieval-augmented generation, longer context windows, tools such as search, and models that reflect on their answers or compare several versions.

Chang then cited Microsoft calling for an AI emergency brake, and Anthropic saying its models could evade shutdown within months. Asked what would convince him the industry should slow down or stop, Huang said agents must be kept in rigorous containment and monitored. If a lab cannot do that, it should not test or deploy them. He said labs with rogue unreleased models should contain and monitor their experiments, not shut down. When Chang said the labs say they cannot contain their models, he called that "made up" and "nonsense". Asked whether he disbelieved the labs, he said he did not know all that they were saying. He said agents that are deployed should get minimum rights, be treated with zero trust and be monitored all the time.

Asked whether Nvidia would stop selling to a lab doing something too dangerous, Huang called it a hypothetical, "but of course". He said Nvidia does not run other companies and its technology is general purpose, but if it knew for certain that a use of its technology was harming people, it would stop.

Cracks in the AI boom and circular financing

During the trip, chip stocks fell for six days in a row, and the narration says CoreWeave, a data centre company Nvidia backs, was hit over concerns about its debt. In Stockholm, Huang spoke at the launch of Sweden's Nvidia-backed sovereign AI supercomputer. The narration says he told reporters that CoreWeave's strength was improving every day. The film says more than a hundred US data centre projects had been blocked or delayed that year. Chang added that AI IPOs were being delayed or scrapped, and asked whether these were cracks in the boom.

Huang said AI is "too fundamental" to go back. Computers used to run software written by hand, and they can now learn problems too complicated for any person to program. He said the industry narrative is "really out of control and ought to be stopped", but that this is an industry growing up, as the internet did. Asked why Nvidia sells so many products, he said it does "as little as we can". Its architecture is general purpose and runs every closed and open model, so it can be both the fastest and the lowest-cost option, like an F1 car that is also a good sedan.

The narration says Nvidia has committed about half a trillion dollars across the AI industry, and critics accuse it of circular financing: giving companies money that they spend on Nvidia products. Huang said AI startups need a lot of computing, so it is not unusual for one to need $10 billion. He said investing connects Nvidia with the best companies, which build on its technology. Nvidia is often an early or anchor investor, but not the lead or largest investor, and owns less than 20% of each company. He estimated that Nvidia has invested close to $100 billion over about five years, which brought in at least five times as much from other institutional investors.

Is dependence on Nvidia a systemic risk?

Chang said most of the leading AI companies depend on Nvidia's hardware and software. Asked whether that dependence is a systemic risk, Huang said AI companies do not depend on Nvidia but choose it. He said Nvidia does not take its position for granted, because it has competitors in every industry and country. When people say America needs to win the AI race, he hears that America's companies should win, not one or two of them.

He said he has no concern that AI can be advanced safely and responsibly at every step. Whether the industry will do so, he said, is "a completely different thing". On supply and demand, he said that at the moment "we're kind of doubling every year" and expects that to go on for several years. He said the infrastructure build-out is about two years old, and earlier technology cycles took about ten years to mature before moving to replacement cycles. He expects the same with AI. If chips become far more efficient, he said, people will do more, not less, because of ambition, and the semiconductor industry will be one of the largest in the world.

Jobs and optimism

Chang asked what he says to someone who learned AI and still sees their job slipping away. Huang called it a theoretical question. He said AI needs manufacturing, so it creates construction, electrician and plumbing jobs, and that $400 to $600 billion had gone into AI startups in the previous six months. He said AI will change everyone's job, and resisting the change will not make it go away.

Asked what happens if his optimism proves wrong, he said he would know he did everything he could, because he is building technology to keep the industry safe. He said people should not talk about the future pessimistically while doing nothing to prevent bad outcomes: "Don't just deliver the bad news. Tell me what you're going to do about it."

How he runs Nvidia, and who comes next

The narration says Huang had led Nvidia for 33 years with no one named to replace him. Chang asked how his management style, with many direct reports and no meetings, had changed as the company grew. He said that style is what allowed Nvidia to grow, because a company with few silos adapts to change. He said a CEO must hold themselves responsible and keep high standards.

Chang said his son and daughter work at Nvidia and asked whether his daughter is in line to succeed him. Huang said his children, Spencer and Madison, would have to earn it. He said the CEO job takes sacrifice, because he is responsible for many families and an entire industry. He said he hoped the next CEO was already at Nvidia, and that he would stay healthy and keep earning his keep. Twenty years from now, he said, he hoped his successor would already be there. He said that just as he can be fired, the next CEO has to be hired.

Weeks after the interview, Nvidia's board accepted his resignation as president, CEO and director. Colette Kress is interim CEO, and the board will consider internal and external candidates for the permanent role^[2].

Sources

1. The Rise and Reckoning of Jensen Huang | The Circuit with Emily Chang — Bloomberg Originals, YouTube (12 November 2026; views checked on 13 November 2026)
2. NVIDIA Announces CEO Transition — NVIDIA Newsroom (12 November 2026)