What Jensen Huang Is Missing: The Harmful Costs of NVIDIA

By C. Rich

Jensen Huang speaks with genuine excitement about the future NVIDIA is building. He describes AI factories that will generate tokens at unprecedented scale, drive economic growth, and solve humanity’s greatest challenges. He sees extreme co-design, scaling laws, and massive compute clusters as the inevitable path forward, a story of progress measured in exaflops and tokens per second per watt.

But there is a side of this story, in my opinion, that rarely appears in his vision.

The same data centers that power these AI factories consume vast amounts of electricity and fresh water, often in communities already facing scarcity and pollution. NVIDIA’s partners are building facilities the size of small cities, and those facilities are not abstract engineering feats; they are real places where local residents compete with supercomputers for drinking water and breathe air made heavier by methane turbines. The environmental cost is not a footnote. It is a direct consequence of the race to scale that Huang celebrates.

The labor required to make these systems possible is equally hidden. Behind every breakthrough in model capability are hundreds of thousands of data annotation workers, many in precarious conditions, labeling the data that trains the next generation of models. These are the people who teach AI to see, speak, and reason, yet they are treated as interchangeable inputs in a supply chain. When Huang talks about the install base and ecosystem, he rarely mentions the human cost of keeping that ecosystem fed with fresh data.

Even the economic story is more complicated than the narrative of abundance. While Huang correctly notes that some jobs will be elevated, the current trajectory shows entire categories of work being hollowed out. Entry-level and mid-tier roles in coding, analysis, customer service, and creative fields are being automated or devalued at a speed that leaves little time for retraining or transition. The new jobs created are often lower-paid, more precarious, and stripped of the meaning that came with the roles they replaced. This is not a smooth evolution; it is a rupture.

Huang speaks of NVIDIA’s success as a computing platform that benefits the entire industry. Yet the concentration of power is unmistakable. One company’s architecture now underpins the AI ambitions of nearly every major player. That dominance allows NVIDIA to shape not only hardware but the direction of research, the priorities of development, and even the public conversation about what AI should be. When a single company sits at the center of the stack, the decisions made in its boardroom ripple across economies and societies.

None of this is to deny the ingenuity or the genuine breakthroughs. NVIDIA’s engineering is extraordinary. The problem is not the technology itself. The problem is the system built around it,  a system that extracts data, labor, energy, and attention at a scale and speed that outpaces our ability to mitigate the harm.

Huang’s optimism is rooted in a belief that the benefits will eventually outweigh the costs. History suggests otherwise. Every previous wave of industrial transformation created real winners and real losers, and the losers were rarely the ones setting the agenda. The AI wave is no different. The question is not whether NVIDIA can build faster and larger factories. The question is whether we, as a society, are willing to accept the human and planetary price of that acceleration without demanding a different path.

The future Jensen Huang describes is technically possible. But it is not inevitable. And it is not the only future we can choose.