Investors may be entering a second phase of the artificial-intelligence trade as enthusiasm begins to spread beyond the companies supplying chips and memory for data centres.
Semiconductor stocks surged in June while Alphabet, Amazon and Meta came under pressure over the cost of building AI infrastructure.
Morgan Stanley strategists now expect some capital to rotate back towards hyperscalers, partly because they have already endured a period of underperformance and may show “more capex discipline in the near-term”.
The bank’s July 6 strategy note did not name the three companies as a new stock basket.
However, separate research from Morgan Stanley internet analyst Brian Nowak provides bullish cases for each as the AI trade broadens rather than abandons chipmakers.
Alphabet turns custom AI chips into a new business
Alphabet is increasingly positioning itself as both a major consumer of AI hardware and a supplier of computing capacity.
Google originally developed its tensor processing units, or TPUs, to run its own AI services.
It is now making that technology available to outside customers, creating another potential revenue stream alongside Google Cloud.
Nowak raised his Alphabet price target to $415 from $375 in a June 29 note.
He estimated that third-party TPU sales could generate as much as $80 billion in additional revenue during 2028 as customers seek alternatives to scarce Nvidia processors and other AI infrastructure.
Improving fundamentals were “creating a tactical buying opportunity for one of the best positioned AI companies around”, Nowak said in the note.
The risk is that Alphabet’s infrastructure spending is also rising rapidly.
The company must expand its data centres before it can sell more cloud capacity and chips, putting pressure on cash flow before the expected revenue arrives.
Amazon offers two routes to AI monetisation
Amazon gives investors exposure to AI through both its cloud-computing operation and its much larger retail ecosystem.
AWS can benefit as Amazon brings more data-centre capacity online and serves demand that capacity constraints previously left unfulfilled.
Nowak believes AWS growth could accelerate beyond 30% if the return on Amazon’s infrastructure investment improves even modestly.
He named Amazon a top pick and retained an Overweight rating and set a $300 price target.
Amazon remained the “most underappreciated gen-AI winner” in Morgan Stanley’s coverage, Nowak said.
Meta’s advertising engine can fund its AI ambitions
Meta is the most direct recovery play of the three because investor concern about its capital expenditure has weighed heavily on sentiment.
“Sentiment has troughed,” Nowak said in a research note reported by Barron’s.
Morgan Stanley continues to view Meta as a top pick and carries a $775 price target, despite doubts about how quickly its enormous AI investment will produce returns.
The company’s existing advertising business provides the foundation for that optimism.
AI already improves content recommendations and advertising performance across Facebook and Instagram, helping Meta increase engagement and give marketers more effective targeting tools.
New products could broaden the payoff. Morgan Stanley identified opportunities including Meta AI search, paid subscriptions, business-focused agents and the possible sale of excess computing capacity.
Nowak estimated that four emerging products could each add between $1 and $3 to Meta’s 2028 earnings per share if adoption meets expectations.
Meta also traded at a lower forward earnings multiple than Amazon and Alphabet when Morgan Stanley named it its preferred large-cap internet stock ahead of first-quarter earnings.
The concern remains execution as investors need evidence that AI can generate advertising, subscription and cloud revenue, not simply a larger capital-expenditure bill.
The post 3 AI stocks Morgan Stanley prefers as investors rotate away from chips appeared first on Invezz