Morgan Stanley remains constructive on AI supply chain and energy security themes, and recommends using energy and value exposures to hedge crowded AI trade risks
AI summary card
Morgan Stanley remains constructive on AI supply chain and energy security themes, and recommends using energy and value exposures to hedge crowded AI trade risks
The report views the recent pullback in Asia-Pacific AI infrastructure as a better entry point after a technical washout, while also recommending allocations to energy security, renewable energy/energy storage, materials, and industrial capex beneficiaries to diversify retail leverage, Middle East developments, and policy uncertainty.
- Morgan Stanley believes computing demand may significantly exceed supply for years to come. While China’s LLM progress poses a competitive threat, it instead reinforces the bullish logic for computing demand under Jevon's Paradox.
- From June 22 to July 17, the MSCI APAC Semis/Equipment Index fell 21%. The report believes this was driven mainly by profit-taking, crowded trades, and the clearing of retail leverage, rather than weakening AI capex fundamentals.
- On positioning, it recommends a barbell structure: AI computing infrastructure on one end, and upstream capex beneficiary themes such as energy, materials, industrials, energy security, renewable energy/energy storage, and defense on the other.
- The Asia thematic watchlist adds Alibaba Group and Aspeed Technology, replaces HD Hyundai Electric with LS Electric, and removes Lynas Rare Earths and Sieyuan Electric.
- Energy security opportunities are quantified as about US$5trn of total investment in Asia Pacific, of which around US$1.2trn by 2030 is incremental spending to reduce import dependence; renewable generation, power grids, and battery/battery-swapping infrastructure are the main value creation areas.
Report interpretation
Overview
This report discusses investment opportunities in Asia-Pacific themes including AI supply chain, energy security, renewable energy/energy storage, and a multipolar world. Morgan Stanley believes that after the recent pullback in AI infrastructure and related themes, the risk-reward has improved, but retail positioning, leveraged products, negative seasonality, Middle East developments, and uncertainty around AI policy still pose short-term headwinds. Accordingly, the report recommends continuing to hold AI computing infrastructure while diversifying portfolios with energy security, materials, industrials, renewable energy/energy storage, and value factor exposures.
Core views
The core views include: first, advances in AI capabilities will continue to boost the value of intelligence and computing power, the economics of enterprise AI use cases remain attractive, and AI capex ROI is still viewed as strong; second, the development of China’s open-weight AI ecosystem intensifies competition but may also expand total computing demand; third, the correction in the Asia-Pacific AI supply chain is more a technical positioning washout than a fundamental turning point; fourth, energy security and the capex supercycle remain long-term themes and can hedge Middle East conflict, energy market volatility, and multipolar policy risks; fifth, regionally, the report continues to prefer Japan and Singapore, while remaining cautious on sectors such as consumer discretionary, staples, and autos.
Analysis framework
The report uses a thematic strategy and cross-industry comparison framework, combining growth, valuation, analyst target price upside, bull/bear scenario risk-reward, thematic relevance, market breadth, fund flows, and positioning signals to screen Asia-Pacific thematic opportunities. For the AI supply chain, the report works with the Asia technology research team to identify 29 key names after the pullback; for energy security, it cites a blue paper to estimate capex scale and the distribution of value creation; for market strategy, it combines factor performance, fund flows, and regional/sector allocation recommendations.
Methodology notes
Rank thematic attractiveness based on growth, valuation, and risk-reward
This framework compares sales and EPS growth across sub-themes, valuations relative to history and across themes, analyst base-case and bull-case returns, and risk-reward skew under bull and bear scenarios, with the average ranking across 11 components forming the overall thematic attractiveness ranking.
Efficiency improvements actually expand computing demand
The report argues that more efficient models and China’s LLM progress will not weaken computing demand, but instead may lower usage costs and broaden application scenarios, thereby reinforcing long-term computing demand.
Pair AI computing infrastructure with energy/materials/industrial capex beneficiary themes
While AI themes remain attractive, near-term crowding and leverage risks are elevated, so the report recommends diversifying with upstream capex beneficiary themes such as energy security, renewable energy/energy storage, materials, industrials, and defense.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- AI computing infrastructureCore preferred theme
- Strengths
- Computing demand is expected to exceed supply for years, AI capex ROI is viewed as attractive, and risk-reward has improved after the correction.
- Weaknesses
- Large prior gains and crowded positioning leave it exposed in the short term to retail leverage and momentum drawdowns.
- Comparison
- Compared with pure momentum exposure, it requires more selective picking of high-conviction supply chain names and pairing with energy and value exposures.
- Risks
- Data center construction bottlenecks, policy and regulatory approvals, US-China competitive actions, and tighter credit markets could delay realization.
- China AI pathwayLeading theme in the ranking framework
- Strengths
- Strong growth expectations, reasonable valuations relative to history and other sub-themes, and the open-weight ecosystem may expand AI applications and computing demand.
- Weaknesses
- Faces US-China policy and market bifurcation risks, and some names may take longer to realize earnings.
- Comparison
- The report says it ranks near the top in the Asia thematic ranking framework and has favorable GARP characteristics.
- Risks
- Policy intervention, model competition, capex cycle volatility, and restrictions on external market access.
- Energy security, renewable energy, and energy storageLong-term theme and macro risk hedge
- Strengths
- Long-term investment is driven by multipolarity, energy shocks, and the need to reduce import dependence; grids, renewable generation, batteries, and battery-swapping infrastructure offer significant value creation potential.
- Weaknesses
- Some sub-industries are already quite well owned, and valuations and order realization still need validation.
- Comparison
- Compared with a single AI exposure, energy security can provide a more direct macro risk hedge and exposure to the capex supercycle.
- Risks
- Middle East developments, energy price volatility, policy execution, supply chain bottlenecks, and technological substitution.
- Japanese and Singapore equity marketsRegional allocation preference
- Strengths
- Japan is supported by reflation, reform, and corporate earnings momentum; Singapore is supported by financials and investment-related sectors.
- Weaknesses
- If global credit conditions tighten or energy shocks intensify, market risk premia may rise.
- Comparison
- The report says Japan remains a core index-level overweight recommendation, and both Japan and Singapore remain core overweight markets.
- Risks
- Rising long-end real rates, slower global growth, energy shocks, and reversals in regional fund flows.
Key data
- MSCI APAC Semis/Equipment correctionDown 21% from June 22 to July 17, 2026The report views this correction as the backdrop for improved risk-reward in AI computing infrastructure.
- Number of key AI supply chain names29Screened by the Asia technology research team using analyst conviction, Top Pick status, target price upside, market cap, and magnitude of correction.
- Factor performance reversalMomentum’s strong year-to-date performance retraced 28%; value and low volatility outperformed by 22.8 percentage points and 19.6 percentage points, respectivelyThe report believes that although momentum rebounded after July 21, it is still too early to conclude that the adjustment has ended.
- Asia-Pacific energy security investment scaleUS$5trn total investment, of which about US$1.2trn by 2030 is incremental spending to reduce import dependenceThe energy security theme is viewed as a long-term capex opportunity and a hedge against macro risks.
- Breakdown of energy security value creationOf US$9trn total value creation, renewable generation accounts for US$2.3trn, power grids US$1.7trn, and batteries and battery-swapping infrastructure US$0.9trnBased on this, the report prefers capex beneficiaries and renewable energy/battery-related exposures.
- Expected sodium-ion battery penetrationAbout 2% in 2027, about 20% in 2030, and about 37% in 2035; global market around 830GWh by 2030CATL is listed as one of the clearest beneficiaries of sodium-ion batteries, while some lithium miners, copper foil producers, and graphite anode companies face substitution risk.
- Changes to the Asia thematic watchlistAdded Alibaba Group, Aspeed Technology, and LS Electric; removed Lynas Rare Earths and Sieyuan ElectricLS Electric replaces HD Hyundai Electric because its order momentum is accelerating faster.
Impact & implications
For investors, the implication of the report is not simply to chase AI, but to use the improved valuation and sentiment window after the pullback to select high-conviction supply chain names while reducing portfolio reliance on crowded momentum trades through energy security, energy storage, grids, materials, industrials, and value factor exposures. In regional allocation, Japan and Singapore remain core overweight preferences, North Asian markets may benefit from a recovery in AI sentiment, but Korea and Taiwan still require close monitoring of retail leverage and the pace of positioning washout.
Risks
- The clearing of retail positioning and leveraged investment products may not be fully complete, which could continue to amplify volatility in Korea, Taiwan, and AI-related sectors.
- Middle East developments may affect energy, transportation, and product markets, and may alter the short-term pricing of the energy security theme.
- Rising long-end real rates and tighter credit market conditions could especially affect tech issuance and capex financing.
- Data center construction bottlenecks could slow the pace of demand pull-through in the AI supply chain and erode pricing power and margins in some segments.
- There is still uncertainty around AI-related policy, including regulatory approvals and US-China competitive actions.
- Although the momentum factor has rebounded, the report believes it is still too early to judge that the technical adjustment has fully ended.
What to watch
- Whether order trends, capex guidance, and data center construction bottlenecks in the AI infrastructure supply chain improve.
- Whether Korean retail margin financing, leveraged ETF fund flows, and short gamma concentration continue to decline.
- Long-end real rates, conditions for tech credit issuance, and broader credit market conditions.
- The impact of Middle East conflict, shipping constraints, and energy prices on expectations for energy security capex.
- US-China AI policy, regulatory approvals, and market bifurcation risks.
- Alibaba Group’s cloud revenue growth and margins, Aspeed’s ramp-up in next-generation BMC products, and LS Electric’s share of data center orders and momentum in new orders.
- Progress in sodium-ion batteries moving from pilot stage to large-scale adoption, and their substitution impact on the lithium, copper foil, and graphite anode supply chains.