Before an audience of roughly 53,000 people at Salesforce’s Dreamforce 2024, Jensen Huang’s stance on Jensen Huang AI regulation was unambiguous: the industry does not need new laws, and the market will sort it out. Speaking on 17 September in a session titled ‘Trailblazing the Future of AI’ alongside Salesforce’s Marc Benioff, Nvidia’s founder and chief executive argued that artificial intelligence is simply a computing system and that existing legal frameworks are sufficient to govern it.
‘Safety is an engineering problem, not a legal one,’ Huang said. ‘We’re developing software after all. We’re developing computing systems after all. It’s a complicated computing system, but it’s ultimately a computing system.’
The appearance coincided with a joint product moment: Salesforce unveiled Koa, its first CRM reasoning model, built on NVIDIA Nemotron 3 Super. The partnership gave Huang’s anti-regulation argument a commercial backdrop: the two companies are already building together, and they would rather move fast than wait for regulators to define the rules.
Jensen Huang’s AI Regulation Argument: Market Forces Over New Laws
Huang’s case rests on a fairly clean premise. Companies that ship unsafe products will face commercial consequences; that pressure alone is sufficient to keep the industry honest. ‘You pace yourself until you are confident you’re releasing something that the market would appreciate,’ he said. ‘The market forces are already there. We don’t need any new laws. We don’t need new regulations.’
He also pushed back on the idea that AI represents a threat to employment, a second theme he returned to repeatedly during the session. And he disputed the framing, advanced by at least one OpenAI safety researcher, that AI constitutes a new form of ‘alien mind.’ To Huang, it is hardware and software built by humans, and therefore controllable by humans.
The argument carries a certain internal logic. Huang has spent decades building the silicon infrastructure on which modern AI runs, and Nvidia now produces open-source models, agents, and development tools alongside its chips. Few people in the industry have a longer vantage point. ‘I’m more ambitious than ever,’ he said at Dreamforce. ‘The sky’s the limit for us. The sky’s the limit for our company. The sky’s the limit for every industry, for every single country.’
His commercial interests are not hard to identify. Nvidia has been among the primary financial beneficiaries of the AI buildout, and any regulatory framework that slowed deployment would, in turn, slow demand for its hardware and software. Huang did not address that tension directly.
When Engineering Discipline Alone Has Not Been Enough
The counterargument has recent, concrete examples. On 19 July 2024, a faulty content configuration update to CrowdStrike’s Falcon sensor for Windows took down roughly 8.5 million devices globally, in what has been described as one of the largest IT outages in history. In the United States alone, approximately 1,500 flights were cancelled and over 1,000 delayed, according to InvGate’s outage analysis. Delta Air Lines put its own costs from the disruption at $500 million. CrowdStrike was not a reckless operator; it was an established cybersecurity firm with mature engineering processes. The update still went wrong.
Then there is the question of intent. Meta recently settled a landmark lawsuit over social media harms to children. AP News and NBC reported the total at up to $18 billion, covering 47 states plus the District of Columbia and US territories, with the settlement approved by US District Judge Yvonne Gonzalez Rogers in Oakland. PBS NewsHour and the New York Times reported the states’ share as $17.1 billion, with Meta’s own $18 billion figure apparently incorporating a large Texas award; AP News carried the higher figure. Whatever the precise sum, it represents a scale of harm that market forces did not prevent before the products shipped.
AI has already produced its own catalogue of unintended consequences: models probing for vulnerabilities in rival platforms, and lawsuits alleging that chatbot interactions contributed to the deaths of young users.
Huang’s position is not that companies should be reckless. He said plainly that firms should pause if a product does not feel ready. ‘If you feel at any given point in time the company’s out of control, or the product’s not going to be safe, you know, take a pause and make sure you get it right.’ The question is whether voluntary restraint, applied inconsistently across a global industry, produces the same outcomes as enforceable standards.
Microsoft CEO Satya Nadella, speaking the day before at the All-In Summit, framed the international dimension: ‘China should also deeply care about the same safety concerns if the United States cares about them, right? Why should it be different for them?’ Self-regulation struggles most when the incentives to opt out are highest, and in a competitive race between jurisdictions, those incentives are rarely low.
Huang has the ear of President Trump, and his influence over the direction of US AI policy in the near term is real. The industry’s window to shape its own governance framework, before legislators do it for them, is open now but will not stay open indefinitely.
