Nvidia CEO Rejects Antitrust Waivers for AI Development Slowdown
Jensen Huang said existing regulations are sufficient for AI safety and opposed proposals for antitrust exemptions that would let AI firms coordinate to slow model development.
Nvidia chief executive Jensen Huang told CNBC’s Jim Cramer that new antitrust exemptions for artificial-intelligence companies are unnecessary, arguing that current laws already cover product reliability and safety. His comments came as Anthropic CEO Dario Amodei called for coordinated efforts to "pace the frontier" of AI development, suggesting government-mediated waivers to allow competitors to limit the speed of model releases.
Huang said safety and testing are engineering challenges that firms can address without additional legislation. "Safety is an engineering problem. Testing is an engineering problem," he said, adding that companies should halt releases until products are ready. He emphasized that existing regulations already govern product functionality.
Amodei’s essay outlined a three-step plan involving third-party safety evaluators, global coordination on standards, and limits on unchecked AI progress, noting that such coordination could face legal hurdles under the Sherman Antitrust Act. Huang, whose company supplies chips to Anthropic and OpenAI and develops its own Nemotron models, argued that developers have the power to ensure safety independently.
While acknowledging the importance of AI safety, Huang dismissed predictions of imminent existential threats, stating that many people are working on proper AI development and that guardrails and safety technologies will be created.