Morgan Stanley: S&P 500 Target at 8,000; Tactically Bullish on HK Internet Stocks
Miles Bennett
Morgan Stanley sets its S&P 500 target at 8,000 points, expects global AI capex to surpass $1 trillion by 2027, and turns tactically bullish on Hong Kong internet stocks with a recommended July-to-September window.
Where are we in the AI super-cycle?
Morgan Stanley says the global AI super-cycle is still in its early stages. The top five cloud providers are expected to spend over $800 billion in capex in 2026, rising to $1–1.2 trillion in 2027.
Across the top 14 cloud providers, annualized capex growth already exceeds 90%. This means → big tech is not experimenting — it is making an all-in collective bet.
Markets worry about "compute overcapacity," but strategist Laura Wang sees it as AI commercialization maturing. OpenRouter data shows first-half 2025 token consumption nearing 35 trillion, up over 15x year-on-year. In plain terms = capacity is being built fast, but it is being consumed even faster.
What supports an S&P 500 at 8,000?
Morgan Stanley targets 8,000 for the S&P 500, projecting 23% constituent earnings growth for full-year 2026.
In H1, capital concentrated in memory, chips, and semiconductors. Hyperscalers lagged because they bore the capex burden. But their valuations have now pulled back sharply, and earnings visibility is strong — investment value is emerging.
H2 should see the rally broaden beyond AI infrastructure. Consumer discretionary, transportation, and regional banks are also favored, driven by tax incentives and capex support from the "Big Beautiful Bill."
Why does Morgan Stanley prefer A-shares over MSCI China?
Morgan Stanley turned fully bullish on A-shares last June. The core reason: tech, advanced manufacturing, and scarce-resource sectors now account for over 40% of the CSI 300, versus under 15% in the MSCI China index.
This means → A-shares offer more direct exposure to China's economic upgrading. Buying the CSI 300 is effectively a bet on "new-quality productive forces."
But A-shares are experiencing a "K-shaped recovery" internally: AI and advanced manufacturing are growing with high certainty, while real estate and consumption drag on traditional sectors. Morgan Stanley does not recommend overweighting consumer or property and stays bullish on AI, semiconductors, and domestic substitution.
How large is the semiconductor localization opportunity?
Wang expects China's chip self-sufficiency rate to rise from over 30% in 2024 to above 70% by 2030. Supply will stay tight long-term, giving the domestic semiconductor chain high-certainty growth potential.
Markets worry that large IPOs like CXMT could strain A-share liquidity. Wang is not overly concerned. This reflects her confidence in policy support — the "national team" has already sold nearly $100 billion in A-share holdings via ETFs and retains the capacity to stabilize the market.
In plain terms = the logic behind domestic substitution is not "can it be done" but "it must be done," and both policy and capital are tilting in that direction.
Why is the HK internet rally only "tactical"?
Wang is tactically bullish on Hong Kong internet stocks, explicitly framing this as a repair window — not a confirmed long-term bull market — and recommends the July-to-September period.
Three pillars support the rally: ① Starting in July, major platforms will release next-generation large models and AI Agent features — catalysts arrive in a cluster. ② Since April, regulators have cracked down on e-commerce price wars, reducing the drag on margins; Q2 earnings will be the key observation window. ③ July lock-up expiries for HK-listed IPOs are being gradually absorbed, easing funding pressure.
After September, however, the market will revert to fundamentals. Whether Chinese internet companies can sustain catalysts in earnings growth, AI innovation, and global market-share gains will determine whether the rally endures.
What does China's rising share of global model usage signal?
OpenRouter data shows Chinese large models' share of global token consumption rose from roughly 5% in early 2025 to 34% by May 2026, surpassing 50% after June.
This means → Chinese models have moved from the periphery to commanding half the global market — far faster than most anticipated.
Wang notes that the post-May share-price pullback in Chinese model companies was driven mainly by concentrated lock-up expiries triggering profit-taking, not by any shift in industry fundamentals. In plain terms = selling pressure pushed prices down, not a deterioration in the business itself.
Content is for reference only, not financial advice.