The AI gold rush is in full swing, but a fund manager is sounding the alarm as he believes the frenzy could turn to dust if the industry doesn't increase its computing power.
Nessim-Sariel Gaon, managing partner of LIAN Group, issued this warning in response to recent market volatility surrounding chipmaker NVIDIA (NASDAQ:NVDA), a major player in the AI sector. Concerns regarding a potential AI bubble bursting have also been voiced by analysts at Goldman Sachs, raising concern about the future health of the AI market.
However, the managing partner says that the true threat to the AI sector is not an inherent bubble, but rather a lack of digital infrastructure – a shortage of data centers specifically designed to handle the immense computational demands of AI. This lack of capacity, the managing partner suggests, could very well cause the bubble to burst.
"Many servers in data centers worldwide are currently inadequate for supporting AI compute," Gaon told Investing.com.
"We must bolster our bespoke digital infrastructure and build cutting-edge facilities to host dedicated GPUs to fuel the AI industry."
Gaon added that the current market enthusiasm for AI, with significant investments from institutions, retailers, and Big Tech, is a positive indicator. They believe the market has strong growth potential but warns that significant compute capacity increase is essential to sustain this growth.
"Investors have to be convinced AI will see long-term sustainable growth," he said. "That's the key to supporting AI's growth. We must plan ahead and build speculatively."
He also contends that the AI bubble is not driven by inflated expectations but by the real risk that the necessary infrastructure for AI innovation will be insufficient. This shortage of servers could result in demand for AI processing far exceeding supply, leading to a market crash.
By rapidly increasing the number of AI-specific servers, the AI sector can ensure long-term stability and avoid the risk of a bubble burst, Gaon concluded.