In an interview with the Financial Times, NVIDIA’s CEO Jensen Huang has firmly rejected the growing comparisons between the current AI euphoria and the dotcom bubble of the 1990s. While many analysts and investors see a parallel to earlier speculative bubbles in the sharp rise in share prices, Huang emphasizes that today’s demand for computing power is real and measurable, in contrast to the oversized infrastructure investments of the Internet era.
During the dotcom era, fiber optic infrastructure was laid on a large scale, far exceeding the demand at the time. A large proportion of these lines remained unused, which in the telecommunications industry is referred to as “dark fiber”, i.e. unlit, inactive optical fibers. At the time, companies had speculated on explosive growth in Int…
In an interview with the Financial Times, NVIDIA’s CEO Jensen Huang has firmly rejected the growing comparisons between the current AI euphoria and the dotcom bubble of the 1990s. While many analysts and investors see a parallel to earlier speculative bubbles in the sharp rise in share prices, Huang emphasizes that today’s demand for computing power is real and measurable, in contrast to the oversized infrastructure investments of the Internet era.
During the dotcom era, fiber optic infrastructure was laid on a large scale, far exceeding the demand at the time. A large proportion of these lines remained unused, which in the telecommunications industry is referred to as “dark fiber”, i.e. unlit, inactive optical fibers. At the time, companies had speculated on explosive growth in Internet use and prepared their networks for a future that never materialized. Jensen Huang now contrasts this historical surplus with the current state of the AI industry: “Today, almost every GPU you can find is in operation,” he explained in the interview. For him, this is proof that the demand for computing power actually exists and is not artificially created.
Source: NVIDIA
Huang sees the current development of AI as a profound technological revolution that goes far beyond chatbots and image generators. Systems are now able to process knowledge independently, categorize research results and act in a context-related manner. While many consumers are only familiar with the visible applications of AI, the industry has long been working on much more complex models and fields of application. As a result, the number of queries and calculations is also growing exponentially, which in turn triggers a direct and sustained demand for GPU performance.
At the same time, Huang recognizes that the industry is facing major challenges. The enormous energy consumption required by the global computing infrastructure could become a central problem. The question of whether cloud providers such as Microsoft or Amazon will be able to efficiently integrate and operate the huge quantities of AI chips also remains unanswered. Nevertheless, it is clear to Huang that the basis of today’s AI wave is based on real demand and functioning use – not on exaggerated predictions of the future, as was the case during the dotcom era.
Conclusion
Jensen Huang does not see the current AI development as a speculative bubble, but as a technologically sound growth phase based on real use and concrete demand. While the dotcom era was characterized by exaggerated expectations and underutilized infrastructure, the AI industry today is characterized by full data centers and growing energy requirements. Nevertheless, it remains to be seen whether the rapid growth is sustainable in the long term or whether the industry is approaching physical and ecological limits.
Source: wccftech.com