A remark made 9 years ago by Jensen Huang (젠슨 황), Nvidia's chief executive, is drawing renewed attention amid a reshaping of the artificial intelligence (AI) industry.

On May 26, IT media outlet TechRadar reported that Huang said in a past interview: "Software is eating the world, but AI is going to eat software."

The comment came in an interview with MIT Technology Review shortly before Google researchers published their paper on the Transformer. At the time, Nvidia was trying to pivot beyond being a simple GPU maker and reposition itself as an AI infrastructure company. The company began actively using GPUs for deep learning training, and the GTC conference, now one of the world's biggest AI events, also began expanding in earnest around then.

At the time, the influence of software-as-a-service (SaaS) companies was growing rapidly. Driven by the shift to the cloud and broader access to computing resources, SaaS firms established themselves at the center of the enterprise IT market. But Huang saw AI as capable of changing how existing software is used and the value chain itself. It meant the structure of the software industry itself could be reshaped, not just through the addition of new functions.

Analysts say there are signs in the market that such a forecast is becoming reality. Each time AI companies such as Anthropic unveil new models, assessments continue that the existing SaaS industry is coming under pressure. The industry is even using the term "SaaSmageddon" to describe AI features encroaching on the existing software domain.

In particular, the spread of agentic AI is cited as a factor increasing anxiety in the software industry. As AI agents become able to carry out some software functions that people previously used directly, the role of software itself could shrink, it says. There are also assessments that AI has already begun to replace some functions in areas such as simple repetitive tasks, information searches and work automation.

The outlet did not, however, conclude that the software industry is collapsing. It said SaaS companies are indeed under pressure, but they have not been completely replaced. Ultimately, analysts say the key depends on how quickly existing software companies can absorb AI and restructure their business models.

Nvidia remains at the center of this change. Once seen as a graphics-card company, Nvidia has now established itself as a core supplier in the AI infrastructure competition. Global big tech companies and startups are competing to secure Nvidia GPUs for AI model training and inference, and the company's AI hardware is spreading rapidly across data centers worldwide.

The industry views this renewed attention as going beyond merely bringing up past remarks again. It says key points to watch going forward will be how far AI can shake up the standing of the existing software industry, and whether software companies can absorb AI to create a new market order.