Anthropic and Nvidia CEOs Clash on AI Safety at Dreamforce
Quick Look
Anthropic CEO Dario Amodei urged slowing AI development for safety at Salesforce's Dreamforce, while Nvidia CEO Jensen Huang argued market forces and self-regulation suffice, days after their debate gained mainstream attention; Huang also confirmed Nvidia-Salesforce partnership on a new reasoning model and referenced a call with Trump about AI data centers.
AI-generated summary
Why It Matters
The AI safety debate has intensified as industry leaders express divergent views on whether voluntary self-regulation or government intervention is needed to manage risks from rapidly advancing AI models.
Anthropic CEO Dario Amodei and Nvidia CEO Jensen Huang shared differing views on artificial intelligence safety during Salesforce's Dreamforce conference, days after the debate burst into the mainstream.
Amodei took the stage first with Salesforce CEO Marc Benioff, during his annual keynote, an appearance that followed an essay Amodei published over the weekend urging the industry to slow the pace of model development.
"I think that's the way to lead the industry forward, to set an example, to say that everyone can always be better," Amodei told Benioff on Tuesday, in front of about 12,000 people at San Francisco's Moscone Center.
Industry executives including OpenAI CEO Sam Altman, SpaceX CEO Elon Musk and Google DeepMind Chair Demis Hassabis, quickly voiced support for Amodei's proposal, which came after industry researchers last week raised the alarm on AI's potential to cause catastrophic harm. Amodei suggested a three-step plan to temper how quickly model capabilities improve without "sacrificing commercial advantage or the United States' lead in AI."
Huang, who appeared with Benioff shortly after Amodei, has a different perspective. The CEO of the world's most valuable company said market forces already exist, and that the industry doesn't need new laws or regulations. Huang added that speed and safety are not mutually exclusive, though companies should pace themselves until they are sure they are releasing something that the market will appreciate.
"It's a false choice," Huang told Benioff. "You could definitely have both at the same time."
Huang said it's up to the model developers to act responsibly with their products.
"Run as fast as you can," he said. "But if you feel at any given point in time the company's out of control or the product's not going to be safe, take a pause and make sure you get it right."
A day earlier, Huang was on stage in Los Angeles for the All-In Summit, where he took a call from President Donald Trump. The president, while on speaker, expressed concern about people who are trying to stand in the way of AI data center construction.
"We're not going to let that happen, sir," Huang said to Trump.
Dreamforce has been one of the premier events in the software industry for well over a decade, attracting about 50,000 people to downtown San Francisco and taking over a large swath of its hometown. Benioff's keynote always includes multiple high-profile guests, often people who are at the center of the tech universe at that particular time.
Benioff has had longstanding relationships with both Huang and Amodei.
On Tuesday, Salesforce and Nvidia announced a new reasoning model for the software company's Agentforce platform, which was built on one of Nvidia open-weight Nemotron models. Late last month, Salesforce and Anthropic expanded their strategic partnership and unveiled "Claudeforce," a new way for salespeople to access critical data directly from inside Anthropic's Claude chatbot.
In terms of how the model companies should proceed, Benioff told reporters on Tuesday, "If they feel like they should slow down, then they should slow down. If they feel they should speed up, they should they should speed up. And then they should be held accountable."
WATCH: Sen. Mark Warner: Have to take advantage of AI leaders calling for regulation
What to Watch
AI outlook — possibilities, not facts
Anthropic will continue advocating for voluntary industry pacing of AI model development
Likely · Within months
Nvidia will maintain its stance that market forces and responsible innovation suffice for AI safety
Likely · Within months
Salesforce will expand its AI partnerships with both Anthropic and Nvidia to enhance Agentforce capabilities
Very likely · Within weeks
Open Questions
- What specific safety measures do Anthropic and Nvidia propose beyond pacing development?
- How will Salesforce's new reasoning model integrate with existing Agentforce and Claude capabilities?
- What are the likelihood and potential form of future AI regulation in the United States?
- Did the call between Huang and Trump result in any concrete policy commitments regarding AI data centers?







