NVIDIA CEO Jensen Huang: AI Doomsday Theory is Nonsense
NVIDIA CEO Jensen Huang stated in an interview with CBS News reporter Jo Ling Kent on September 19 that the claim that "2030 will not be the end of the world" has a "0% chance" of becoming reality, directly refuting recent pessimistic views in the industry about AI potentially leading to extinction-level risks.
In the interview, Huang further stated that the industry should "push forward with AI technology as quickly as possible, regardless of any other factors." He described predictions that AI could lead to the end of human civilization as "without scientific basis," continuing his previous public rebuttals of the "AI doomsday theory."
Notably, in August, during an interview at the All-In Summit, Huang expressed a position that was clearly different from this one: while he similarly stated that "doomsday predictions have no scientific basis," he also acknowledged that "if companies feel they are out of control, they should pause or slow down their R&D pace," and publicly praised whistleblower Jacob Coxon for showing "great courage" in exposing internal safety concerns at Anthropic. In contrast, his current statement has shifted from "companies should self-regulate their pace based on circumstances" to "the industry should push forward at full speed," indicating a tightening of his stance against slowing down.
During the same event in August, Trump unexpectedly called Huang and publicly stated, "robots will not take over the world, AI will not take over the world, this whole thing is a scam"; although Huang did not fully agree with this statement at the time, he did express the need to ensure that everyone in the U.S. wins in the AI race. Analysts believe that Huang's recent tightening of stance is closely related to the Trump administration's inclination to accelerate AI development, as well as NVIDIA's own business interests in relaxing export controls (especially regarding chip exports to China) and maintaining an unregulated AI development environment.
Huang's latest statement comes at a sensitive time as OpenAI and Anthropic prepare for IPOs—despite Anthropic repeatedly emphasizing AI safety issues and CEO Dario Amodei's long-standing calls to slow down AI development, Huang chose to use the almost absolute term "0%" to publicly oppose the cautious voices within the industry regarding safety.
From an industrial interest perspective, as the head of NVIDIA, the largest supplier of AI computing power globally, Huang's business interests are highly tied to the overall expansion speed of the AI industry—faster AI development leads to greater demand for computing power, which in turn increases NVIDIA's chip sales; conversely, any form of industry self-limiting or tightening regulation could directly suppress the growth space of its core business. Therefore, Huang's continued downplaying of AI risks and encouragement for the industry to move forward at full speed is objectively creating a favorable public opinion environment for the continuous expansion of his core business, which contrasts with the "safety first" statements made by some model manufacturers due to their own reputational and regulatory pressures.
The report also mentioned that Huang previously commented that some cybersecurity alerts are essentially "creating market panic to stimulate demand," and questioned, "who wouldn't want their market to be enthusiastic about their products?" This consistent rhetoric is also seen as a continuation of his tendency to downplay various technological risk warnings.
Source: Public Information
ABAB AI Insight
Huang has long held a skeptical or even opposing view towards the "AI risk theory"—from his early public statements that "AI safety alerts are overhyped," to comparing some industry safety concerns to "marketing tactics to create demand," and now using the term "0% probability" to outright deny the existence of extinction-level risks from AI. His public statements show a clear path of continuously downplaying risks and emphasizing accelerated development. In August, he made a rare statement that "companies should pause R&D based on their level of loss of control," but a month later shifted to "should push forward at full speed," indicating that his stance is rapidly adjusting with changes in the external political and business environment.
Huang's statements are highly tied to NVIDIA's core business interests—as the largest supplier of AI computing chips globally, NVIDIA's revenue and stock price growth directly depend on the continuous investment in AI computing power across the industry; any form of "slowdown theory" once accepted by the market could suppress customer capital expenditure expectations, thereby impacting NVIDIA's own orders and valuation. Therefore, Huang's choice to continuously release signals of "acceleration is reasonable, risks are controllable" in public is essentially using his influence to create a more favorable public and policy environment for the continuous expansion of his core business, which is highly consistent with his previous efforts to relax chip export controls to China.
A comparable historical case is the typical divergence that occurs in emerging technology industries during their expansion phase, where "suppliers downplay risks, while regulators and some practitioners call for caution"—for example, the public divergence between early equipment manufacturers in the nuclear energy industry and some scientists regarding safety risk assessments essentially reflects the upstream equipment and resource suppliers, who, due to their own business interests, tend to downplay the systemic risks that downstream applications may bring. The current AI industry is at a split stage where leading labs (such as Anthropic and OpenAI) are racing for record IPO valuations while emphasizing safety and caution, while computing power suppliers publicly advocate for accelerated development, making the internal positions around the core issue of "development speed" increasingly public and politicized.
This incident essentially reflects the struggle for discourse power regarding "development pace" within the AI industry chain: model manufacturers (especially those emphasizing safety issues like Anthropic) are trying to strengthen the risk narrative to justify potential self-restraint or regulatory intervention, while computing power supplier NVIDIA is attempting to deny the existence of risks to maintain an unregulated environment for continued capital expansion. Mechanically, which narrative ultimately dominates will directly determine whether future capital and policy resources continue to flow unreservedly towards computing power expansion or shift towards safety governance and risk review—this is why Huang chose to publicly and prominently deny AI doomsday risks at the sensitive juncture of OpenAI and Anthropic both racing for IPOs, as this struggle for "narrative dominance" is fundamentally related to the future flow of trillions in capital.
ABAB News · Cognitive Law
- Those who profit from acceleration will deny the necessity of slowing down.
- A stance that changes with the wind indicates that the stance has never been principled.
- The dominance of narrative determines the flow of capital.