Trump and Nvidia's Jensen Huang publicly reject calls to slow AI as a Chinese ploy

A call timed to a slowdown debate

As hani.co.kr reported in a column by freelance writer Lee Jae-woon, the idea of putting a brake on AI development was gaining attention in September when Donald Trump phoned into an industry event in Los Angeles. He appeared there by phone in a public conversation with Nvidia CEO Jensen Huang. He dismissed warnings about AI risk as a scam and argued against the idea of slowing development.

The venue matters. California is home to Silicon Valley, and the event was a place where the tech industry speaks for itself. The columnist calls it unusual for a US president to join such a gathering by phone, and reads it as a sudden decision. Some see it as Trump matching a plan Huang had already drawn up.

Who started the slowdown talk

According to the column, the first major industry figure to raise the idea recently was Dario Amodei, CEO of Anthropic, the company behind the Claude assistant. He did so the day before Trump's call. His main argument was safety: AI is advancing so fast that humanity lacks the time to build protections.

The column says OpenAI's Sam Altman and Elon Musk, who runs SpaceX and whose AI is called Grok, also seemed to agree. The writer finds it notable that the people who gain most from fast AI progress were the ones urging restraint, and says the remarks should not be taken at face value.

He offers two possible readings, both his own interpretation. One is that the executives wanted to show society they take safety seriously. The other is that they wanted to build international pressure on China, which seems unconcerned with slowing down.

Trump's 'Chinese ploy' argument

Trump's reply was aimed at China, the rival in the race for AI leadership. He said that efforts to slow AI development stem from political motives or a Chinese scheme. The columnist treats this as more than conspiracy talk. In his view it is evidence that the current US administration believes this logic and acts on it.

He adds a political reading. Trump cannot afford to look behind China in AI, and with the Iran issue already a headache, he benefits from stressing the threat from an adversary while showing the US ahead. That also fits what Silicon Valley wants to hear.

Huang's incentives differ from Trump's

The column traces Nvidia's rise to its early push to use graphics chips (GPUs) for the heavy calculation behind AI, work once done by general-purpose processors (CPUs). It also credits a close partnership with South Korea's SK hynix on high-bandwidth memory, a fast type of memory that AI chips depend on.

Tension between Washington and Beijing has, in the columnist's telling, effectively closed off sales of Nvidia's newest products to China. That leaves Huang needing strong demand from the US, South Korea, India, Europe and elsewhere. The columnist sees the call as an event staged to serve that need.

The two men point the same way but want different endpoints. Huang wants US-centered AI to stay ahead while Nvidia's GPUs remain the top choice everywhere, including in China. Trump wants the US to keep AI leadership and, the writer suggests, to hold Nvidia within its sphere of influence. These are the columnist's inferences, not statements from either man.

The lunch and what comes next

Later in September, at a White House lunch, Huang and Amodei reportedly debated the issue briefly. Huang is said to have asked, in the column's words, "Why make extreme warnings in public?" Amodei reportedly replied that model capabilities should be stated honestly and risks not underestimated.

The columnist reads both remarks through interests. Amodei stresses how capable Anthropic's models are, and the company is pursuing a Nasdaq listing in November. Huang is telling people not to chill AI growth with needless worry.

The writer expects the exchange may remain a passing episode, since China shows little interest in slowing down and most US opinion does not favor it. Still, he predicts the idea will gain some recognition and responses from governments and companies. He sees one wild card: physical AI, meaning AI with a body or machine to act in the world, which could make things harder to predict.

This is an opinion column, so treat its motive-reading as one person's analysis. The reported facts are who spoke, when and what position each took.