To avoid the AI bubble bursting, the number of bespoke AI machines in data centers worldwide will have to increase tenfold by 2026, says leading AI expert and managing partner of LIAN Group Nessim-Sariel Gaon.
LIAN Group is an investment firm that builds and funds successful companies across multiple industries, including digital infrastructure, AI, healthcare and blockchain. Since launch, they have deployed over $500 million of invested capital.
Gaon’s intervention follows volatility in chipmaker and sector titan Nvidia’s share price. Its recent instability has been coupled with predictions that the AI bubble might burst from the likes of Goldman Sachs, raising concern over the future health and growth of the market.
While he recognizes the concern over the future of the AI sector, especially after its boom in 2022, Gaon believes the real threat to its vitality is its lack of capacity – its lack of digital infrastructure.
In its current state, Gaon suggests this shortage in capacity could lead to the sector’s bubble bursting.
Nessim-Sariel Gaon, managing partner of LIAN Group, said: “Ever since ChatGPT’s breakthrough in 2022, and even the surge in NVIDIA’s share price, commentators all across the industry have warned about the AI bubble bursting.
“But, beyond all the fearmongering, let me tell you, the threat is very real. And it’s capacity that will be the knife in the coffin.
“Given the amount of data and processing power needed to accommodate AI programs, a lot of the servers globally are just not fit for purpose. This is the real ceiling on the expansion and growth of the AI sector; it’s a real barrier for innovative AI startups looking to disrupt the market.
“The number of bespoke AI machines – and by that, I mean the servers able to accommodate reams of AI data – has to increase if we’re going to avoid a slump. In fact, they’ll have to increase tenfold and certainly by 2026. At the end of the day, it’s a ticking time bomb. Quick action is needed.”
