Morgan Stanley sets S&P 500 8,000 target on AI capex cycle

The S&P 500 can reach 8,000 by year-end as AI infrastructure spending defies fears of a slowdown, according to Morgan Stanley's China equity strategist.

Morgan Stanley set an 8,000 year-end target for the S&P 500, projecting 23% earnings growth in 2026 as the AI investment cycle enters a "very early stage" that will sustain capital expenditure above market expectations.

"The AI super cycle remains in its infancy, and capital spending will continue to exceed consensus," Laura Wang, chief China equity strategist at Morgan Stanley, said in an interview with Tencent Finance.

The bank estimates the top five global cloud providers will spend more than $800 billion combined in 2026, rising to as much as $1.2 trillion in 2027. Among the top 14 hyperscalers, annualized capital expenditure growth has exceeded 90%. Morgan Stanley raised its 2026 and 2027 CapEx forecasts by 5% and 17% respectively after the latest earnings reports. Token consumption reached nearly 35 trillion in the first half of 2026, up more than 15-fold from about 2 trillion a year earlier, according to OpenRouter data.

The call comes as the Philadelphia Semiconductor Index enters a technical bear market, down more than 20% from its peak, with Wall Street divided on whether the selloff signals a structural shift or a buying opportunity. Morgan Stanley's U.S. equity strategist Mike Wilson said chip stocks may bounce but are unlikely to reclaim market leadership, favoring hyperscale cloud operators instead.

Wall Street divided on chip stock bottom

The semiconductor selloff has been the fastest and deepest in recent history. The SOX index tumbled 10% in a single week, its worst performance since April 2025, pushing cumulative losses past 20% from a record closing high set a month ago. Memory chip stocks have fallen 36% from their peak, while U.S. AI tech beneficiaries are down 25%, according to Goldman Sachs data.

JPMorgan strategist Mislav Matejka's team argues the pullback reflects sector rotation and momentum trade unwinding rather than deteriorating fundamentals, with the Relative Strength Index approaching oversold territory. Goldman Sachs' Mark Wilson said the momentum factor drawdown — 28% peak-to-trough in 17 trading days versus a historical median of 22% over 33 days — is "nearing its end" but lacks a near-term reversal catalyst.

Morgan Stanley's Wilson expects market gains to broaden beyond technology. The bank favors consumer discretionary and transportation stocks as potential new drivers, alongside hyperscale cloud companies that have stronger core business support and room for cost optimization.

China tech and the tactical HK rebound

Wang also turned bullish on Chinese equities, citing the CSI 300's more than 40% weighting in technology, advanced manufacturing and scarce resources — sectors that benefit directly from the AI cycle. She described the current period as a tactical rebound window for Hong Kong-listed internet platforms, driven by three factors: a密集 pipeline of new AI model launches from July, easing price competition in e-commerce after regulatory intervention in April, and the gradual absorption of IPO lock-up expirations.

Chinese large language models have rapidly expanded their global footprint. OpenRouter data shows Chinese models accounted for 5% of global token consumption in early 2025, rising to 34% by May 2026 and exceeding 50% after June. Wang said the third quarter will be a critical window for Chinese internet companies to demonstrate whether their AI investments translate into commercial returns.

On the semiconductor front, Wang expects China's chip localization rate to rise from above 30% in 2024 to more than 70% by 2030, driven by demand growth that outpaces domestic capacity expansion. She said overseas investors have shown "high interest" in upcoming large-scale A-share IPOs including ChangXin Memory Technologies, and that the national team retains capacity to stabilize markets after selling nearly $100 billion in ETF holdings.

This article is for informational purposes only and does not constitute investment advice.