Goldman remains bullish on AI as cloud investment target nears $1.2 trillion

Goldman's analysis zeroes in on Asian AI hardware markets, including Taiwan, South Korea, and mainland China, all of which are on holiday today.

These Asian markets rely on sustained cloud spending plans even as bond yields keep rising. With cloud spending forecast to reach $1.2 trillion by 2027, investors have a rationale to buy on dips rather than reduce holdings during rate-driven declines. The note's key insight might be the timing: markets could stay volatile into early November, then potentially rally by year-end. That timeline coincides with the US midterms, Tuesday's White House AI meeting, and a week when Chinese markets are closed.

This report, along with Morgan Stanley's barbell call and UBS's constructive stance, reinforces the broad consensus among major banks that the AI cycle still has momentum.

---

Despite climbing bond yields, Goldman holds its AI conviction. The bank contends that $800 billion in cloud spending combined with Asian valuations near historic lows outweighs the rate headwind, and it predicts that volatility around the midterms will precede a year-end rally.

Here are the key points:

• Timothy Moe of Goldman describes AI stocks as still appealing even with higher government bond yields.
• Cloud hyperscalers are projected to spend roughly $800 billion this year and approximately $1.2 trillion by 2027.
• These outlays are a crucial demand indicator for Asia's AI hardware supply chain.
• Asian stocks are trading at about 10 times earnings, near the bottom of their historical range.
• Corporate earnings growth is expected to cushion the impact of high interest rates.
• Moe predicts choppy markets until the US midterms, then a potential year-end rebound.

According to Jinshi Data, Goldman Sachs continues to find AI stocks valuable even as government bond yields spike. Timothy Moe, the firm's Asia-Pacific equity strategy head, highlighted hefty cloud spending as a strong backing for the region's hardware supply chain. Moe noted that the top cloud computing firms are set to invest about $800 billion this year, potentially reaching nearly $1.2 trillion by 2027, a demand factor he considers vital for Asian AI hardware producers.

These remarks arrive as the global bond rout challenges investor enthusiasm for growth equities. Long-term Treasury yields have risen to roughly two-decade highs this week, increasing the discount applied to future profits and weighing on high-valuation sectors. Moe contended that corporate earnings expansion should act as a shield against the high-rate backdrop.

Valuation forms another leg of his argument. Asian shares are trading at around 10 times earnings on average, close to the lower end of their historical band, which Moe termed very low. This affords the region greater capacity to withstand rate pressure compared to markets with already extended valuations, and it bolsters the case for AI-related shares.

Moe does not anticipate a trouble-free near term. He predicts markets will stay somewhat turbulent ahead of the US midterm elections in early November, as elevated energy prices and geopolitical uncertainties add strain. After that period ends, however, he foresees room for a rally by year-end, fueled by earnings growth and a rebound in valuations.

This perspective aligns with a wider change in how big banks are positioning on AI. Instead of ringing the death knell for the theme, strategists more and more contend that the investment cycle still has legs, even as they discuss where the best bets are. For Asia, which hosts a large portion of the AI hardware supply chain, the magnitude of intended cloud expenditure stays the foremost metric.

The danger to that scenario is that bond yields continue to ascend and eventually compel the largest spenders to reconsider their investment tempo. For the moment, Goldman wagers that the dedication to AI infrastructure, paired with low valuations throughout the region, will surpass the headwind from elevated rates.