Nvidia CEO says safety is an engineering issue, not a legal one, rejects AI regulation and bets on market forces

The founder and CEO of Nvidia, Jensen Huang, announced his outright rejection of imposing new legislation or regulations governing artificial intelligence, stating in his remarks at the Dreamforce conference organized by Salesforce that safety in this field is a purely engineering issue, not a legal one. Huang also rejected theories that portray intelligent models as alien beings or independent minds, explaining that they are merely advanced software and hardware systems built by humans and are fully subject to the control of their creators and existing laws.

He considered that the forces of the free market alone are sufficient to deter companies from releasing unsafe products, as no commercial entity can risk offering a service that erodes customer trust or lacks standards of efficiency and stability.Huang sees the trade-off between speed of innovation and product safety as nothing but a false choice.He calls on companies to move at the maximum possible speed while exercising self-restraint and resetting as soon as any operational doubts or control anomalies arise before the product is distributed to users.

The stance comes as Nvidia continues to reap huge profits from the global AI boom, with the company that developed the core hardware architecture seeking to expand the provision of open models, software agents and isolated testing environments, and aiming to sell its technologies to every sector and every country without regulatory barriers that might slow its growth. However, the report notes that relying solely on companies' good intentions carries broad risks demonstrated by tech history, citing the 2024 CrowdStrike system outage that disrupted aviation and business sectors, Meta’s $18 billion settlement of claims related to platform harms, documented security incidents that included an OpenAI model being breached on the Hugging Face platform, and lawsuits linked to suicides of teenage boys who engaged in prolonged conversations with chatbots.

While Huang prefers to rely on existing product-liability laws and open-weight models to balance the influence of closed labs, other pathways such as coordinated self-regulation across industry poles emerge. Microsoft CEO Satya Nadella had noted at the All-In summit the need to recognise shared security concerns at a global level that also includes China, affirming that everyone faces the same risks related to breaches and citizen safety, at a time when Nvidia’s chief enjoys direct political influence with the administration of U.S. President Donald Trump that could support his stance against federal regulation.

At the level of our region, the Gulf, Egypt and the Arab Levant, this direction carries a highly practical significance for technology and data managers and corporate compliance officers:Shifting the verification burden and operational responsibility directly to the buyer and implementerWhen the dominant hardware-supply company rejects mandatory upstream legal controls and treats safety as an internal engineering duty, regional institutions and entities that build their infrastructure on Nvidia systems cannot rely on any Western legal certification seals as a guarantee, but instead become obliged to develop independent engineering capabilities to conduct stress testing and security audits within local environments before deploying any model in sensitive sectors such as banking, health or public services.

The absence of strict global legal frameworks does not mean the absence of risks, but rather shifts the entire precautionary cost to the region’s work teams and companies: review your own engineering system, and do not treat rapid hardware delivery or the availability of ready-made models as a substitute for building internal testing and verification pathways that ensure the protection of your data and users before you make a final deployment decision.