After crashing and burning for two months due to investor wariness over inflation in the United States and fears of an artificial intelligence bubble, AI related stocks are once again climbing the global stock markets, according to a Bloomberg report.
On Friday, Japanese and South Korean stocks posted among their best weekly gains, where the chip-heavy Nikkei 225 Stock Average advancing 4.7% for the week, Topix extending its record high run, and the KOSPI registered 22% jump in 10 days due to the AI chip rally, the publication added.
Hitoshi Asaoka, chief strategist at Asset Management One told Bloomberg, “A huge amount of hyperscaler money is flowing into hardware. That is translating into extremely strong sales and profit growth for hardware companies.”
This is a reversal from the steep losses since July, after major tech player’s earnings reports showed massive AI budgets for uncertain outcomes and refuelled concerns about overspending and a market bubble.
What Elon Musk’s post indicates about rise in AI stocks
After weeks of shaky ground, AI related stocks are on the rise again, and amid this, billionaire Elon Musk engaged in discourse on his social media platform X (formerly known as Twitter), that could explain the resurgence in investor confidence in AI.
But will this rally sustain? Elon Musk‘s post indicates that it will. Or it at least provides an insight into what tech stocks AI could push further. Responding to a post by entrepreneur Peter Diamandis on X, where he wrote: “Memory, not compute, is the rate limiter of the Agentic Era.”, Musk replied: “Few realize this” (sic).
The above posts indicate that as AI tasks evolve from more simple routines into complex “agentic” roles, demand for memory and storage — which is already high — will further skyrocket and this is barrier is currently being underestimated.
Notably, the demand is projected to increase to levels creating a “shortage” and with chip producers factoring in backlog in their documents already. Said producers of memory chips — namely SanDisk, SK Hynix, Samsung and Micron — have all led the recent rally.
Will the AI-backed rally for chip stocks continue?
SK Hynix, which last month made history as the biggest public listing by a foreign company in the US stock market, said the AI boom has “redefined the memory chip business for good”. CEO Kwak Noh-Jung told Bloomberg that customers are increasingly looking for long-term supply agreements and believes the shortage situation is likely to extend further than 2030.
Kwak noted that in the past, the industry has been plagued by cyclical supply gluts, and “ups and downs. But things have clearly changed.” The ongoing AI plans back Kwak’s optimism. Trillions of dollars are being poured into data centres — Oracle, Microsoft, Meta, Amazon and Alphabet (Google), have added a cumulative debt near $350 billion to build hardware for their AI ambitions.
SK Group Chairman Chey Tae-won told Bloomberg TV he believes that demand for chips is likely to outrun supply till artificial general intelligence is achieved aka the point at which AI systems are generally smarter than humans.
Kwak thinks, this can take decades — similar to the growth of the internet, saying: “We spent almost 30 years finishing the completion of the internet infrastructure. When it comes to AI, I believe that the AI industry size is much, much larger than the internet.”
Supply bottleneck the key challenge
The Bloomberg report noted that while competition from China is a factor, near-term demand “outlook remains intact”. Qian Zhang, emerging markets equities investment specialist at Baillie Gifford told the publication that limited supply will be what poses a challenge for the sector.
“Because of AI agents and physical AI, memory demand has exploded, but we entered into this with a quite limited supply capacity — that’s where the bottleneck is. We’re not saying the world will be building data centers forever at this speed, but that is a real physical bottleneck that only a few companies in this world can resolve,” Zhang stated.
(With inputs from Bloomberg)
