How is the AI bubble "blown up" by tech companies?
Analysis suggests that cloud computing giants, by investing in AI startups and requiring them to develop models on their own cloud platforms, not only increase the revenue of the cloud platforms but also inadvertently amplify the valuation of chip companies. Currently, large cloud platforms have supported a significant portion of NVIDIA's data center revenue. Once this investment cycle ends, NVIDIA's growth engine may be severely impacted
There may be a huge bubble in the AI field, which has become a basic consensus on Wall Street. Most people believe that investors' frenzy has caused this bubble. In contrast, well-known corporate strategy consultant Drew Breunig offers a completely different perspective.
In an article published on Monday, Breunig specifically mentioned the role of large cloud computing companies in the AI bubble.
These cloud computing giants have huge cash reserves and have invested billions of dollars in companies building foundational models. More importantly, they require these companies to develop models on their cloud platforms, creating cloud computing revenue while also investing.
Subsequently, they use their newly acquired cloud computing revenue to purchase large quantities of high-performance GPUs and rent them to AI model building companies and other market participants. This large-scale purchasing inadvertently inflated the valuation of GPU manufacturers, further exacerbating the bubble in the entire AI industry.
Breunig believes that this model is precisely what distorts the valuation of the AI market. He points out that the income obtained by cloud platform companies through investment is fundamentally different from the valuation method of traditional venture capital, and this difference to some extent amplifies the bubble in the AI field.
For example, large cloud platforms have supported a significant portion of NVIDIA's data center revenue. Once this investment cycle ends, NVIDIA's growth engine may be severely impacted.
The article also presents several scenarios that could break this bubble:
Investor pressure, a decrease in the valuation of startups, and efficiency improvements brought about by technological progress could all be factors in breaking this bubble.
If investors begin to question the cost-effectiveness, energy consumption, and reliability of AI technology, or if startups run out of funds before finding clear application scenarios, even cloud platforms may struggle to sustain this bubble.
Breunig also mentioned the potential impact of technological progress on the bubble. He believes that as AI models become more efficient, the demand for computing resources may decrease, which could reduce the need for high-performance GPUs and affect the entire ecosystem