Wall Street has found AI’s new choke point

SK Hynix and Micron crossing the $1 trillion valuation mark shows Wall Street has already moved into the next phase of the AI trade — and many investors are still positioned for the last one.

This is the warning from Nigel Green, CEO of global financial advisory giant deVere Group, as SK Hynix and Micron both surged past the $1 trillion valuation threshold for the first time on Tuesday as booming demand for AI memory chips continued reshaping global markets.

Semiconductor stocks helped drive world equities to another all-time high, while US stock futures advanced and European futures traded mixed.

“Wall Street spent the past two years believing Nvidia was the AI trade,” Nigel Green says.

“Now the market is starting to realise the real power may sit deeper in the supply chain.”

The explosive rise in memory-chip makers signals a major shift taking place beneath the surface of global markets.

Investors are no longer simply rewarding companies building AI products.

“Increasingly, they are concentrating capital into the firms controlling the infrastructure bottlenecks the entire AI economy depends on.

“And memory has rapidly become one of the biggest.”

Without advanced high-bandwidth memory chips, AI systems cannot operate at the speed or scale tech giants require. Every major company racing to dominate artificial intelligence — from Microsoft and Meta to Amazon and Google — is now locked into an escalating battle for compute power, semiconductor access and memory capacity.

Demand is accelerating so aggressively that markets have started treating memory less like a commodity and more like a strategic asset.

“The AI boom is becoming less about who builds the models and more about who controls the choke points,” explains the deVere CEO.

“Savvy investors are now seeing the leverage.”

For years, memory-chip manufacturers were viewed as volatile cyclical businesses vulnerable to oversupply and collapsing margins. AI has completely changed the economics.

Now investors see companies like SK Hynix and Micron sitting at the centre of one of the largest infrastructure spending waves in decades.

Tech giants are collectively spending hundreds of billions of dollars expanding AI data centres and compute capacity. Investors want exposure to the companies supplying the hardware underneath that expansion.

“The market has started aggressively repricing the infrastructure layer of AI,” Nigel Green notes.

“And investors crowding into the same AI names they bought 18 months ago are increasingly looking in the wrong place.”

This concentration of capital is now driving broader markets.

A relatively small cluster of semiconductor and AI-linked stocks is responsible for an outsized share of equity gains globally, pushing indices higher even as other sectors struggle with weaker growth and pressure on earnings.

“Semiconductor stocks are no longer simply participating in the rally. In many cases, they are the rally.

“The assumption built into current valuations is that AI spending keeps accelerating for years.

“That may ultimately prove correct. But investors should understand how dependent markets are becoming on a very narrow part of the global economy.”

The comparison increasingly being made across trading desks is oil.

Not because memory resembles energy, but because access to AI infrastructure is rapidly becoming essential to economic power, corporate competitiveness and national security. Governments are already treating semiconductors as strategic assets while corporations unable to secure enough compute capacity risk falling behind rivals.

At the same time, concerns are growing over the wider consequences of artificial intelligence.

“Investors are betting AI transforms the global economy,” Nigel Green says.

“But markets are also making another bet — which is that the companies controlling the infrastructure underneath AI become more valuable than the companies building the products people actually see.”

He adds: “The market has already moved on from the first phase of the AI trade.

“A lot of investors still haven’t.”