Morgan Stanley's Chief China Economist, Robin Xing, shared fresh perspectives during a closed-door meeting this week, focusing on how to balance the dual priorities of technology and domestic demand.

From a liquidity standpoint, Xing reiterated the firm's view that the Federal Reserve will not raise rates this year, and may even cut them next year if necessary. He clarified that the market's anticipated balance sheet reduction is not equivalent to tightening. Kevin Warsh's proposed "shrinking" is a reduction in tools, not liquidity—it will not tighten overall macro-market liquidity.

On AI capital expenditure, Nvidia recently partnered with six top financial institutions to launch a $500 billion independent computing power financing platform, addressing the sharp deterioration in overseas cloud providers' free cash flow—which could hit -$140 billion by 2027, creating a massive funding gap. The platform brings in third-party capital, isolates circular transaction credit risk, and Nvidia provides residual value guarantees in exchange for profit-sharing above breakeven—essentially a financial innovation born of necessity.

Xing believes that structured assets with underlying collateral, such as ABS and CMBS for computing centers—especially high-quality computing equipment collateralized debt—offer better risk-reward profiles than unsecured, highly-rated corporate bonds.

For China, he suggested drawing lessons from U.S. experience: consider launching computing power asset securitization, special bonds, and long-term capital (social security, insurance funds, local quasi-fiscal resources) to support high-residual-value domestic computing nodes, while exploring accelerated depreciation tax incentives to optimize tech firms' cash flows and boost domestic AI hardware adoption, gradually raising localization rates.

Meanwhile, as global weather grows more extreme, there will be greater emphasis on energy security and climate-resilient infrastructure investment. Europe needs to upgrade power grids, expand wind power, and even improve regulations to install air conditioning in response to heatwaves. China, in turn, can expand ultra-long-term special treasury bonds and local special bonds to fund underground pipeline renovations and flood control infrastructure along the southeastern coast.

Chief Strategist Laura Wang noted that the early-July recommendation to shift investment focus to Hong Kong has been validated, with the Hang Seng Index and MSCI China Index rebounding over 13% and nearly 10%, respectively. However, at this mid-August juncture, she believes the risk-reward balance for Hong Kong has shifted. She advises gradually reallocating positions back into A-shares, and not increasing risk exposure to Hong Kong further.

Hong Kong stocks have a negative correlation with overseas liquidity. After a significant correction in U.S. equities—especially AI-related sectors—markets have now stabilized. Additionally, Morgan Stanley upgraded Korea to an overweight position nearly two weeks ago, forecasting KOSPI to reach 9,000 by mid-next year, implying roughly 30% upside. Funds previously flowing into Hong Kong may return to Japan and Korea, creating selling pressure on Hong Kong.

In the short term, A-shares are seeing more pronounced improvement in conditions, warranting priority attention. Investment Daily (liulishidian) has curated the key takeaways from Morgan Stanley's closed-door meeting below.

Overseas Financing Innovation: Nvidia's $500 Billion Platform

We are now in mid-August, the earnings season for domestic and international companies, and also a few weeks of dense macroeconomic data releases. There's considerable focus on how to break the K-shaped divergence in the domestic economy. At the same time, we're watching overseas developments—U.S. inflation, shifts in Fed policy paths, and especially the innovative financing methods emerging for AI, which carry both insights and risks.

Today's theme centers on balancing technology and domestic demand. Ultimately, we must answer: in the second half of AI investment, what will markets primarily focus on? There are overseas references worth considering—first, the Fed's policy trajectory; second, the novel financing techniques emerging in the AI sector.

For example, last week, Nvidia, along with several top financial and asset management firms, launched a $500 billion new platform for financing. This was driven by near-term pressures: overseas cloud providers' free cash flow is deteriorating sharply. The major players' free cash flow could turn negative to the tune of -$140 billion by 2027, creating a huge funding gap.

This is when Nvidia and six institutions set up the $500 billion independent computing power financing platform. On the surface, it introduces third-party capital, isolating the credit risk of circular transactions that had been criticized, with some improvement in the business model. Notably, Nvidia provides residual value guarantees to reassure investors, in exchange for profit-sharing above the breakeven line. Essentially, this is financial innovation—though born of necessity.

Shrinking the Balance Sheet Isn't Tightening: Room for Rate Cuts Next Year

Many worry that if AI requires trillion-dollar financing, will it make the sector highly sensitive to Fed policy? We've repeatedly emphasized over the past months that we don't expect the Fed to raise rates this year. Recent July CPI and PCE data were moderate, confirming that the Fed likely has confidence to hold rates steady, and even has room to cut next year if needed.

Of course, some still worry about balance sheet reduction even if rates stay put or fall next year. After all, incoming Fed Chair Kevin Warsh keeps mentioning an upcoming balance sheet reform. But we believe the market is misreading this—shrinking the balance sheet does not equal tightening. Warsh's "shrinking" is a reduction in tools, not liquidity tightening. These are two different things.

What's Warsh's framework? It would lower the Treasury General Account balance, but simultaneously reform the tiered reserve rate mechanism, guiding banks to buy short-term Treasuries. This relies on passive maturity of short-term debt to achieve the "tool reduction." In essence, it maintains ample reserves within this framework, without tightening overall macro-market liquidity.

Thus, the shrinking process won't transfer long-duration risk to private credit or AI financing markets as feared, and its impact on actual financial conditions would be minimal.

Computing Power ABS, CMBS Collateralized Debt Even Superior to Unsecured High-Grade Bonds

That's why our fixed income and private credit teams still believe that, in the U.S. credit market, structured assets with underlying collateral—such as ABS and CMBS for computing centers, especially high-quality computing equipment collateralized debt financing—offer a reasonable risk-reward profile. Compared to unsecured, highly-rated corporate bonds, these collateralized debt instruments backed by premium computing assets are slightly more attractive.

China's Computing Power Financing: Lessons from U.S. Experience

What does this mean for China? China also faces risks of low-price deflationary cycles, with room to further optimize liquidity through RRR cuts and rate reductions—macro liquidity cannot be reduced. On the other hand, what policy tools are in reserve? Especially if August-September economic data shows no clear improvement, three structural themes—computing power financing, climate adaptation "dual-purpose" projects, and consumption support—could be considered for policy reference.

For instance, compared to various U.S. financing innovations, China could consider policy optimizations and innovations in AI computing infrastructure, from financing to tax incentives, to address constraints on tech capital expenditure. Could we introduce special investment and financing support tools for AI computing power, such as asset securitization? Could we encourage local coordination of long-term capital—including social security funds, insurance capital, and local quasi-fiscal resources—while using special bonds to provide debt and equity financing support for domestic computing nodes with high residual value and high utilization rates? This would gradually catalyze domestic AI hardware adoption and raise localization rates.

Now, I'll hand over to our Chief Strategist Laura to analyze how we view U.S. and Chinese capital markets given the AI investment landscape and Fed policy expectations.

Morgan Stanley's Chief China Strategist Laura Wang: Gradually Rotate Positions Back to A-Shares, No Further Hong Kong Exposure

There's quite a bit to share this week, including adjustments to our views on Chinese equities. In May, we repeatedly emphasized our belief that Hong Kong stocks would have a strong rebound window starting in July, and market developments have largely matched our expectations. Since July, the Hang Seng Index and MSCI China Index have rebounded over 13% and nearly 10%, respectively.

Now in mid-August, we believe the risk-reward balance for Hong Kong has shifted. At current levels, we recommend gradually rotating positions back into A-shares and not increasing risk exposure to Hong Kong further. This is tied to global capital market linkages. Our first implicit assumption: has the significant global market adjustment since late June largely cleared, moving into a relatively stable, even clearer outlook? Our answer is yes.

The factors that caused volatility have been absorbed over the past period. As mentioned, our view on the Fed under Kevin Warsh's leadership regarding balance sheet reduction: we expect it could begin as early as Q1 next year, lasting about two years, with a reduction scale of $1.5 trillion. But throughout this process, we believe the Fed will maintain abundant liquidity on the supply side. The capital markets shouldn't experience obvious liquidity scarcity or tightness.

We also forecast no rate hikes this year, with two cuts next year. Overall, the impact on capital markets and risk appetite should be relatively manageable.

Funds May Flow Back to Japan and Korea: Hong Kong Faces Selling Pressure

Second, after the significant correction in U.S. equities—especially AI-related sectors—indices tied to extreme risk appetite have gradually stabilized. Additionally, we believe Japan's reflation story and the continued narrative of rising ROE remain strong beyond AI. So, we maintain a very optimistic view on Japan.

We also upgraded Korea to overweight nearly two weeks ago. While KOSPI has rebounded over 20% from its low, it still has nearly 30% upside to our target of 9,000 by mid-next year. Over time, we expect gradual fund flows into Japan and Korea. In this process, Hong Kong, as a source of funds, may face increasing relative selling pressure.

So, we formally suggest: if you followed our early-July advice and allocated heavily to Hong Kong, there's no need to further increase exposure at these levels. For stocks with substantial gains, consider appropriate adjustments.

Observing Hong Kong Market: Signals to Watch from Mid-September

How do we view Hong Kong's trajectory going forward? In the second half, we see some Hong Kong-specific opportunities that could re-solidify momentum or even boost performance. But timing matters. We've always emphasized global capital flows—especially the high liquidity of financial assets, which move in and out quickly.

Entering September, especially mid-to-late September, we should closely watch whether the following factors could re-emerge and push Hong Kong's investability back to a relatively high level.

First, global market volatility has a negative correlation with Hong Kong. Could geopolitical risks and U.S.-China relations ease? Based on various announcements, especially from the U.S. government, there's a schedule for a visit to the U.S. on September 24. Before and upon confirmation of the official visit, could there be signs of easing on trade or tech restrictions? This would help reduce Hong Kong's risk premium.

Second, many major internet companies listed in Hong Kong are likely to heavily promote their AI advancements—new feature launches, new model releases, and so on. The next round of catalysts could unfold over the coming month or so, including WeChat's Agent function, Qwen's Model 4, and new releases from Zhipu and MiniMax. We think these could impact markets after September—we'll wait and see.

Third, if fiscal stimulus emerges on consumption, it would be a major positive for both Hong Kong and A-shares. Hong Kong has new consumer plays, autos, and many internet e-commerce names that would benefit significantly from increased fiscal policy support.

If these three signals appear, we believe Hong Kong could see a new wave of market conditions driving another rally after September. But for now, we recommend focusing more on A-shares in the short term. As large-scale IPOs are gradually digested, we believe A-share conditions will improve more noticeably.