Investing in the stock market requires authoritative, professional, timely, and comprehensive analysis to uncover potential thematic opportunities.
On July 17th, at a mid-year investment strategy conference hosted by China Asset Management, renowned scholar and advisor Liu Yuhui provided a detailed analysis of the triggers behind the current market's high volatility and the trajectory of the AI sector.
He posits that three cyclical dimensions—geopolitics, monetary-financial cycles, and the AI industry—will dictate the market's future direction.
Liu Yuhui believes the current global market is experiencing extreme divergence, with a massive influx of capital being siphoned towards the AI and semiconductor sectors.
The US Philadelphia Semiconductor Index has been in a roaring bull market, with gains of up to 90% in this single cycle.
However, it has recently begun to exhibit intense volatility, which he views as a normal development.
This reflects a shift in the US monetary and financial cycle; the US has clearly entered a period of dollar tightening.
The dollar's tidal cycle is shifting to a receding mode, increasing the strain on global US dollar liquidity.
Liu Yuhui explained how the US has historically leveraged successive dollar tidal cycles to reap global benefits, and this instance is no exception.
Of the US stock market's $83 trillion market capitalization, nearly $30 trillion is held by foreign capital.
The comprehensive interest cost on the $39 trillion US national debt is approximately 3.7% or even higher.
The challenge for policymakers is to simultaneously extend the duration of US Treasuries to lower interest rates while supporting the high-duration US stock market.
The only viable path may involve releasing high volatility, as recently seen in the sharp swings of the Philadelphia Semiconductor Index, which embodies this dynamic.
From the perspective of the AI industry cycle, Liu Yuhui believes it may have just begun.
However, capital markets are experiencing dramatic sentiment swings, potentially hitting the peak of a bell curve.
In the short term, there is a sharp conflict between AI producers and consumers—upstream players are profiting handsomely while downstream players are bleeding cash.
Hardware gross margins are as high as 70%-80%, while the free cash flow of the five major US cloud providers is being depleted.
This state of affairs is unstable and will inevitably lead to capital market fluctuations as the issue demands resolution.
Regarding AI's future, he stated that China possesses the world's most comprehensive industrial system, capable of continuously breaking down technological barriers.
"The ultimate destination of AI will belong to China."
Finally, Liu Yuhui emphasized that remembering two key principles is essential for achieving excess returns in this market.
First, in the hierarchy of China's macro-policy decision-making, great power competition takes precedence, followed by technology and the modern industrial system, with domestic demand ranking last.
Second, safety now takes priority over efficiency.
The global economy has become "fortified" and "weaponized" as a fact.
Regardless of the process, the era of low-cost globalization dividends has ended; safety premium is now, and will be, the most significant asset premium.
The following are the key insights from Liu Yuhui's presentation, as summarized below.
The Market's Extreme Divergence
If we talk about industries, globally—at least from a capital markets perspective—it seems there is only one industry: semiconductors.
All the world's capital is surging and being siphoned in this direction.
Consequently, we see deflation, declines, and even bear markets elsewhere.
Only AI infrastructure and the AI supply chain, represented in the US by the Philadelphia Semiconductor Index, are in a roaring bull market because all the global capital is being drawn in.
Since February and March of this year, gold has turned bearish, and so has Bitcoin.
China's capital market, the A-share market, appears extremely polarized, but the same is true globally, including in the US.
The rise in US indices is essentially driven by the Philadelphia Semiconductor Index.
Since April, the Philadelphia Semiconductor Index has surged by up to 90% in this cycle.
However, it has recently begun to release intense volatility, which is quite normal.
Most US stocks, including the "Magnificent 7," have not risen this year.
The leader among the Magnificent 7, the AI supply chain's anchor—NVIDIA—has reached a $5 trillion market capitalization.
But how long has it stayed at $5 trillion? It has been stagnant at this level for ten months without moving.
During these ten months, how much has the Philadelphia Semiconductor Index risen? 90%. That's today's market.
This dynamic is mirrored in China's A-share market as well.
This year's market is extremely polarized; all capital is being absorbed into semiconductor hardware, while everywhere else is in a bear market.
Geopolitical Pressures Resurface
How will the current situation evolve in the future?
From my research perspective, I believe three main dimensions will determine the future course.
The first is geopolitics.
The second is the monetary and financial cycle.
The third is the AI industry's own cycle.
It has been nearly three years since the emergence of ChatGPT at the end of 2022, which formally ushered AI into its industrial cycle wave.
As humanity's last great industrial revolution, where does the industry cycle stand after three years?
What sharp conflicts exist? How will it evolve? We examine this from these three dimensions.
First, geopolitics. At this current juncture, the geopolitical landscape appears fragile.
While global oil prices remain stable, this stability relies on rapid crude inventory drawdowns over the past four months, now nearing critical levels.
Whether capital markets will absorb and price this risk remains uncertain.
Second, the Russia-Ukraine conflict. The balance has shifted, tilting towards Ukraine.
Thus, geopolitical pressure is rising again.
The backdrop for China today has only one script: great power competition.
Everyone knows that on this global stage, only two players remain; others are on the menu.
These two are like participants in a prolonged sports match, with the score alternating.
This dynamic national scoreboard reflects the ebb and flow of geopolitical interests and pressures between the two sides.
Why are Chinese assets under significant pressure this year?
Looking through a geopolitical lens, one can see it's essentially the great power competition, with the underlying scoreboard subtly shifting. This is the first factor.
The US Strategy: Harvesting the Globe via Dollar Tides
The second factor is monetary and financial dynamics, which is likely the most influential among the three factors currently at play.
The US has clearly entered a period of dollar tightening; there is no doubt about it. The global supply of US dollars is contracting, and liquidity is tightening.
Observe how all dollar-denominated assets are facing a liquidity crunch: US Treasuries are falling, commodities are falling, cryptocurrencies are falling, gold is falling, precious metals are falling.
If the liquidity shortage worsens, the pressure will inevitably spill over to today's hottest assets: semiconductors and AI hardware.
Because further liquidity needs would require liquidating some of today's most buoyant assets to reduce financial leverage.
Recent volatility reflects this reality: the degree of the global US dollar liquidity shortage is rising again.
As it rises, market participants will inevitably liquidate some of the most buoyant assets. Hence, the recent volatility in the Philadelphia Semiconductor Index reflects this larger monetary-financial cycle.
For the US, its current strategy is an open play.
The US's core competitiveness today is dollar hegemony.
How does it harvest global geopolitical benefits to solve its own, seemingly insurmountable problems? It relies on the pricing power of its monetary and financial system—dollar hegemony.
The US's open strategy is to use the pricing power of the international dollar system to attract global capital to the US, transforming it into today's AI computing infrastructure: chips, data centers, energy storage grids, IDCs, ultimately forming AI's physical infrastructure.
This physical infrastructure remains in the US.
But who bears the brunt of the asset price bubbles created in global markets? All the families around the world. That is dollar hegemony.
Today, the US stock market is a behemoth. At its peak before the recent adjustment, its total market capitalization reached $83 trillion.
This $83 trillion is different from before; the proportion held by foreigners is at a record high, possibly close to $30 trillion.
This capital has flowed in over the last year or two, including money from the Middle East and the massive foreign exchange surplus created by China. Capital from all over the world continues to flow into the US.
Next, it will use dollar hegemony and the dollar's tidal cycle to reshuffle the deck repeatedly through the release of high volatility.
After releasing high volatility and driving the market down, the Federal Reserve steps in to buy the dip, repeating the process many times.
This could involve many small, frequent tidal cycles or one large cycle.
After several cycles, people will eventually find that the ownership of these computing infrastructure assets, after being washed back and forth, ends up in American hands.
This is the game the US plays.
Moreover, over the past 80 years since the post-war period in 1945, the US has repeatedly used this tidal cycle to consolidate its dollar hegemony and resolve its internal cost issues.
Looking globally, can the US succeed again in this round of the dollar's tidal cycle?
The Variable for Breaking the Pattern: Dismantling Barriers with Industrial Power
I believe there is likely only one variable today: the presence of a major Eastern power.
The global game is different now; it's a G2 great power competition.
Only a major power can break this predetermined pattern; only a major power retains the ability to prevent the dollar's tidal cycle from succeeding.
Because the incentive mechanism of Wall Street's current wealth system is to "protect technological barriers."
Through its deep capital markets, Wall Street protects innovation, protects technology, and attracts the world's smartest money and largest capital.
But for the rising Eastern power, the core value of its rise is precisely "dismantling barriers."
Because China has built the world's most comprehensive industrial system, the most complete and powerful global industrial supply chain capability.
We know China has its "DeepSeek moment"; in fact, this is just a snapshot.
I believe that over time, we will witness the birth of one DS moment after another, showcasing China's industrial hard power to break technological barriers and reshape industry structures.
Your model may be strong, backed by massive capital creating a huge computing model, but China possesses powerful engineering optimization capabilities that can instantly drive down token costs.
In reality, the world's need for top-tier models is minimal, less than 10%.
90% of tasks can be solved with first-class models.
China, with its multitude of first-class models, standing on the shoulders of giants and leveraging its super-strong industrial efficiency optimization capability, is seeing a rise of numerous first-class models.
This means the capital invested by the US rapidly becomes China's own AI computing infrastructure.
What the US is left with is the future burden of massive capital expenditure depreciation. This is China's capability.
The Policymaker's Dilemma: Balancing Stock and Bond Durations
The US is also formidable. Its economic management team has ended internal strife, becoming a powerful entity comprised of three individuals.
We often hear about potential conflicts or games between the US economic spokesperson and policymakers. This issue is now resolved.
They are fellow disciples, with their mentor being Druckenmiller. They form the US's powerful economic iron triangle, working to manage this situation.
What is the inherent contradiction within the US financial system today? It's the dollar and US Treasuries.
The dollar and US Treasuries are two behemoths. US national debt stands at $39 trillion with a comprehensive interest cost of about 3.7% or higher.
Additionally, the US today has capital expenditures for AI, attracting significant credit inflows from shadow banking, which also increases debt and raises costs.
Simultaneously, the other behemoth is the US stock market, valued at about $80 trillion, with nearly $30 trillion belonging to foreigners.
Therefore, the policymakers' dilemma is finding a balanced approach to maintain a high duration for both these massive asset classes simultaneously.
Both the economic spokesperson and the policymakers strongly desire to extend the duration of US Treasuries, lower interest rates, and create conditions for rate cuts.
But this is a challenge because no one has the power to simultaneously support both behemoths at a high duration.
A trade-off between the two seems inevitable.
In a sense, there may be only one path: to release high volatility and divergence, thereby lowering the high duration of the elevated US stock market.
This could create conditions for extending the duration of US Treasuries and lowering interest rates.
Hence, the recent high volatility seen in the Philadelphia Semiconductor Index and US stocks reflects the intense, often unseen,博弈 beneath the surface from a monetary-financial cycle perspective.
AI Industry Cycle in its Infancy, but Markets May Be Topping
From the perspective of the AI industry cycle, I personally believe it may have just begun.
As humanity's greatest and last industrial revolution, it still has a long way to go.
However, after three years, several internal contradictions are now coming to a head.
The first core contradiction is the intense conflict emerging between AI producers and AI consumers.
AI producers are the semiconductor companies.
Represented by semiconductor chips, a range of AI hardware is currently enjoying peak红利 due to various shortages.
Capital expenditures have disrupted supply-demand balances, leading to shortages in electronic components, memory, optical connections, etc., with gross margins普遍 at 70-80%.
On the other hand, the consumer surplus from AI is being maximally squeezed.
Today's AI consumers are represented by the five major US cloud providers.
Their free cash flow is being burned through, forcing them to borrow from Wall Street.
But borrowing rates on Wall Street are high, with the 10-year Treasury yield at 3.5%, creating another significant矛盾.
The current state is extremely polarized and unstable.
This issue must be resolved, and its resolution will cause capital market fluctuations.
Note that any great technological revolution follows a technology maturity curve, like the Gartner Hype Cycle.
This curve is actually a合成 of two curves.
From an industry cycle perspective, progress should be steady.
However, the capital market's emotional cycle reflecting this industrial revolution is dramatic.
Thus, we see a bell-shaped curve in the middle before stabilization. This bell shape reflects sentiment, not a physical process.
The curve is an overlay of two lines: one representing the心理/炒作 curve and the other representing the actual industry growth curve.
Their叠加 creates the波涛汹涌 capital market映射 of the industry cycle. Simply put, the波动 is significant.
What is the approximate current state? I believe it's likely at the peak of the bell curve's头部.
From an industry cycle standpoint, the AI industry cycle may have just begun.
But from a capital market experience perspective, we may be冲击 a顶部.
Therefore, investing in this great AI industrial transformation requires a firm grasp of this curve's规律.
Triggers for High Volatility Stem from Two Main Areas
To summarize, today's global AI industry cycle is influenced by three intertwined factors: the monetary-financial cycle, geopolitics, and the AI industry's own cyclical规律.
The current fundamental重心 lies with the monetary-financial cycle.
The reason for the recent release of volatility fundamentally stems from dollar tightening; the dollar's tidal cycle has clearly entered a receding mode.
Europe, Japan, Australia, and Canada have raised rates; a US rate hike is inevitable.
Conditions are gradually成熟, partly due to geopolitics.
For China, as a key geopolitical player, we are strategically and deliberately weaponizing the supply chain, exporting China's global supply chain dominance.
In economic terms, this is the deliberate use of industrial power to export inflation abroad, albeit at the temporary cost of enduring weak domestic demand.
Thirdly, from the AI industry cycle perspective, internal factors triggering high volatility mainly stem from two areas.
In the short term, AI has entered a vicious cycle where producers and consumers are in a尖锐矛盾期; upstream is profiting effortlessly while downstream is bleeding.
AI today faces an extremely steep cost slope alongside a very flat revenue slope.
The entire US AI industry chain's profits rely on five cloud providers supporting the entire semiconductor hardware industry—this is an open secret.
The long-term challenge for AI is resolving the contradiction between productive forces and production relations, as mentioned earlier.
As the last industrial革命 replacing intelligence, it fundamentally alters the endogenous production function of human economic activity.
Originally, our production function had only three factors: capital, labor, and technology.
Now, an exogenous variable has appeared: the AI entity created by humans.
This intelligent entity, as an exogenous variable, disrupts the function entirely.
We imagine a uniform K-shape recovery, but what actually emerges is massive wealth concentration in the hands of the world's smartest, most powerful capital that masters AI first,占有 the entire蛋糕.
The resulting macro structure carries the潜在 risk of总需求萧条.
If not addressed, without designing a合理的治理结构, how we confront and承载 this great human AI industrial革命 remains a long-term命题.
These issues, both short and long-term, will be priced into capital markets, bringing intense volatility.
AI's Ultimate Destination Belongs to China
I believe that following the evolution of industrial规律, the ultimate destination of AI will belong to China.
This industrial revolution differs from the internet revolution.
The internet era was about software吞噬一切.
Today's AI era has a capital density and intensity 100 times greater or more than the internet revolution.
The surface表象 of AI is models, algorithms, and算力.
But behind it lie data centers, thousands of interconnected GPUs, with a single IDC's power consumption equivalent to a small-to-medium city.
It is supported by an entire heavy-asset industrial system—what we call the HALO ecosystem.
To make the AI industry run requires electricity, chips, storage, servers, connectivity, cooling, and a series of精细化工 materials industries.
Looking further, as AI evolves into physical AI—entering终端侧, real-life scenarios like robots, smart cars, smart factories—it becomes inseparable from various mineral resources.
Mineral resources become crucial, but what truly matters is the ability to extract and refine them.
This ultimately映射 the underlying industrial capability, which resides in China, in the Eastern power.
Therefore, the final outcome of future AI competition depends on a强大的 industrial system and capability, which恰恰 exists in China.
Thus, the deterministic前景 and guiding direction of this AI industry cycle most likely belong to China.
Hierarchy in Policy Prioritization: Safety Over Efficiency
There are two principles I've emphasized for over a year. Remembering them can help you navigate this cycle over the next five years and smooth out volatility.
The first principle is to remember that China's current macro-policy decision-making follows a specific hierarchy of priorities.
We must not be distracted by noise. The top priority is undoubtedly great power competition, as that is the only script for this era.
Second is technology, and third is the modern industrial system. These are the weapons and means.
Only fourth comes daily necessities, our domestic demand.
Therefore, in China's A-share market, to achieve excess returns, one must clearly understand this priority weighting when investing.
The second principle is to understand that the world cannot go back. The era of globalization's low红利 and efficiency is over.
The true state of today's world can be described in eight characters: "Safety takes priority, efficiency steps back."
The global economy has become highly fortified and weaponized; this is a fact.
Consequently, what will be the most richly溢价 asset in the world going forward? The safety premium.
This must be firmly kept in mind.
The function and responsibility China's capital market is meant to undertake have been clearly defined.
Therefore, the market's narrative, pricing structure, and valuation structure must adapt to this change.
To quote an official stance: "When the focus shifts from how much money was made in the past to how difficult the problems that can be solved in the future are, the underlying logic of China's capital market has undergone a qualitative change. This is also an直观体现 of the Chinese economy转向 innovation-driven."
This is the official tone. One must truly understand this to effectively navigate China's capital market, to "Be water, my friend," achieving the highest good like water, where the adaptable prevail.
Comments