China's Growth Awaits a Fiscal Impulse as Sovereign AI Capital Expenditure Expands Alongside Wider Credit Spreads

Overview

This report discusses, in sequence, China's growth and fiscal execution, sovereign AI capital expenditure, AI-related debt financing, and US long-term Treasury yields. Its core view is that an improvement in China's near-term growth depends on whether the remaining fiscal impulse can be deployed promptly; sovereign AI will create long-term infrastructure opportunities but will also entail technological restrictions, execution challenges, and wider credit spreads.

Core views

Regarding China's economy, the report notes that real GDP growth in the third quarter is currently tracking at approximately 4.4%Y, once again below the growth target. Deflation and fiscal gridlock remain the primary constraints, while the economy's K-shaped divergence is also deepening. The report views the fiscal impulse as the most critical near-term swing factor: as of the end of July, around Rmb2trn in fiscal and quasi-fiscal capacity remained unused for August through December, and the pace of its release will directly affect growth performance for the year. Slow fiscal execution is primarily evident in two areas. First, general bond financing has not been as significantly front-loaded as it was in 2025. Second, local government special-purpose bond issuance has been slow due to factors including tighter reviews of local projects and a still-insufficient pipeline of projects related to the "Six Networks." Government bond issuance accelerated in the second half of August, but the report's reasoning is that the remaining quota does not itself equate to effective demand support; the key issue is whether it can be converted into spending and project implementation. Sovereign AI is the report's second main theme. Economies are seeking greater control over AI infrastructure, data, and supply chains, thereby promoting localized procurement, sovereign computing capacity, cloud services, semiconductors, and AI capability development, while in some cases bringing in trusted international partners. This strategy affects not only chips and models but also propagates along the value chain to data centers, power infrastructure, software and cloud services, colocation data centers, robotics, and AI applications. The capital expenditure figures provided in the report indicate the considerable scale of this cycle: US AI capital expenditure in 2026 is estimated at US$860B, while China's related capital expenditure from 2026 to 2030 is estimated at RMB35tn. Morgan Stanley believes that the trend toward sovereignty will reinforce the AI capital expenditure cycle while increasing the likelihood of strategic technological restrictions. Fragmentation of the technology ecosystem may not suppress long-term growth; instead, it could accelerate domestic AI adoption and extend AI capabilities to the "Global South." The report regards sovereign AI infrastructure as one of the fastest-growing opportunities globally, with a particularly positive view of construction demand associated with data centers and cloud services, while emphasizing that capital deployment capacity and actual execution remain the primary obstacles. AI development also requires substantial debt financing. AI-related issuance currently accounts for approximately 7% of the investment-grade index. The report believes that, at current levels, credit markets still have ample capacity to absorb additional supply. However, supply will not enter the market without cost; wider spreads will serve as the adjustment mechanism, attracting capital through higher credit compensation and balancing issuance pressure. Regarding US interest rates, the report believes that the primary driver of the recent rise in long-term Treasury yields and steepening of the curve is Treasury repricing associated with energy prices and the Federal Reserve's reaction function. Investment-grade corporate bond supply may also have had some impact, but the report judges that its effect is neither persistent nor significant. Therefore, recent curve movements should be understood primarily in terms of macro interest-rate expectations rather than a corporate bond supply shock.

Analysis framework

The report first assesses China's growth conditions using third-quarter GDP tracking estimates and economic divergence, and then analyzes whether the remaining fiscal impulse can be deployed based on budget execution and the issuance pace of general bonds and local government special-purpose bonds. It subsequently maps the capital expenditure impact of sovereign strategies along the AI infrastructure value chain and illustrates the scale of the cycle using investment figures for the United States and China. Finally, it evaluates financing pressure through the share of investment-grade bonds, market capacity, and the spread adjustment mechanism, before decomposing the macroeconomic and credit-supply factors behind the steepening of the US Treasury curve.

Methodology notes

• Macroeconomic framework

  Fiscal Impulse and Budget Execution Tracking

  The report combines unused fiscal and quasi-fiscal capacity, the pace of government bond issuance, and the actual deployment of expenditure to assess the marginal support from fiscal policy for near-term growth.

• Industry/Sector Analysis FrameworkUpstream, Midstream, and Downstream Value-Chain Transmission

  Sovereign AI Value-Chain Transmission

  Starting from the development of sovereign computing capacity, semiconductors, and cloud capabilities, the report analyzes how capital expenditure transmits to power, data centers, software and cloud services, colocation facilities, robotics, and AI applications.

• Fixed Income and Credit AnalysisSpread analysis

  Additional Bond Supply and Credit Spread Adjustment

  The report measures the scale of supply using the share of AI-related bonds in the investment-grade index and considers wider spreads the primary balancing mechanism through which the market absorbs additional issuance.

• Fixed Income and Credit AnalysisYield curve analysis

  Attribution of Long-Term Treasury Yields and Curve Steepening

  The report distinguishes the effects of macro-driven Treasury repricing from investment-grade corporate bond supply and concludes that repricing associated with energy prices and the Federal Reserve's reaction function is the primary cause of the recent curve steepening.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Chinese Economy and Fiscal Impulse

  The remaining fiscal and quasi-fiscal capacity is a key swing factor for growth performance from August through December.

  Strengths

  Around Rmb2trn in policy capacity remains unused, and government bond issuance accelerated in the second half of August.

  Weaknesses

  Third-quarter real GDP growth is tracking at only 4.4%Y, while deflation, fiscal gridlock, and K-shaped divergence persist.

  Comparison

  General bond financing has been less front-loaded than in 2025.

  Risks

  If special-purpose bond issuance and actual spending continue to be constrained by project reviews and an insufficient pipeline, fiscal support may struggle to translate into growth promptly.

• Global Sovereign AI Infrastructure

  Localization, sovereign computing capacity, and supply-chain control will reinforce the AI capital expenditure cycle and affect data centers, cloud services, semiconductors, power, and AI applications.

  Strengths

  The report views it as one of the fastest-growing infrastructure opportunities globally, particularly in data centers and cloud services.

  Weaknesses

  Capital deployment capacity and project execution remain the primary obstacles.

  Comparison

  US AI capital expenditure in 2026 is estimated at US$860B, while China's is estimated at RMB35tn from 2026 to 2030.

  Risks

  Strategic technological restrictions may tighten, while ecosystem fragmentation will also increase construction and supply-chain complexity.

• AI-Related Investment-Grade Credit

  Expanding AI capital expenditure will create additional demand for bond financing.

  Strengths

  The report believes that, at current levels, the market still has ample capacity to absorb additional issuance.

  Weaknesses

  Increased supply will require wider spreads to attract capital.

  Comparison

  AI-related issuance currently accounts for approximately 7% of the investment-grade index.

  Risks

  Concentrated issuance may cause credit spreads to widen further.

• US Long-Term Treasuries

  The recent rise in long-end yields and steepening of the curve primarily reflect macro interest-rate repricing.

  Weaknesses

  Energy prices and changes in the Federal Reserve's reaction function are creating repricing pressure at the long end.

  Comparison

  Although investment-grade corporate bond supply has had some impact, the report believes its effect is neither persistent nor significant.

  Risks

  If energy prices and expectations regarding the Federal Reserve's policy response continue to change, long-term yields may remain volatile.

Key data

• Third-Quarter Real GDP Growth Tracking Estimate4.4%YThe report states that current growth is once again below target
• Unused Fiscal and Quasi-Fiscal ImpulseRmb2trnAs of the end of July, expected to be available for use from August through December
• US AI Capital ExpenditureUS$860B2026e
• China AI Capital ExpenditureRMB35tn2026-30
• Share of AI-Related Issuance in the Investment-Grade Index~7%The report believes the market currently still has ample absorption capacity

Impact & implications

The report's overall implication is that an improvement in China's near-term growth depends not only on the fiscal quota but also on government bond issuance, the project pipeline, and the efficiency with which funding is converted into actual expenditure. Global sovereign AI development could create a sustained infrastructure capital expenditure cycle and drive demand for data centers, cloud services, and related supply chains. Financing markets still have capacity, but additional supply may require wider credit spreads, while changes in US long-term interest rates are primarily driven by energy prices and expectations regarding the Federal Reserve's policy response.

Risks

• Deflation and fiscal execution gridlock in China may continue to weigh on growth, while tighter reviews of special-purpose bond projects and an insufficient project pipeline could delay policy implementation.
• Sovereign AI development may face stricter strategic technological restrictions.
• Capital deployment and project execution for sovereign AI continue to face significant obstacles.
• Increased AI-related bond supply may require wider credit spreads to be absorbed by the market.

What to watch

• Monitor the actual pace at which approximately Rmb2trn in fiscal and quasi-fiscal capacity is released from August through December.
• Monitor whether accelerated government bond issuance translates into spending and whether the pipeline of local government special-purpose bond projects improves.
• Monitor the share and volume of AI-related investment-grade bond issuance and changes in credit spreads.
• Monitor the impact of energy prices and the Federal Reserve's reaction function on US long-term Treasury yields and the shape of the curve.