Letter to our Co-investors 1Q25

Theme and Background

The chapter opens by noting that the market in 2025 has presented two distinctly different phases: the first phase saw a clear weakening of the U.S. market, while the second phase witnessed a global market decline triggered by the U.S. government's announcement in April of the most stringent tariff policies since the Great Depression of the 1930s. The report argues that this event has far-reaching implications and therefore devotes significant篇幅 to analyzing the tariff policy and its impact.

Core Thesis

The author's core investment argument is: Do not easily bet against the decline of the United States, but the current extreme concentration and high valuations in the U.S. market already constitute a significant risk. The report challenges the market consensus that "the U.S. stock market is always the best long-term investment," arguing that the relative underperformance of the U.S. market in the first quarter of 2025 has begun to shake this belief.

Counter-Intuitive Judgments:

• The S&P 500 recorded its worst-ever relative performance against the European STOXX 600 in the first quarter of 2025, while Chinese equities rose over 10% during the same period.
• The author believes the "invincibility myth" of the U.S. market may be ending, but this is not a recommendation to directly short the U.S.

Key Arguments and Data

1. Historical Performance Comparison (May 2020 to end of 2024, EUR-denominated including dividends):

Index	Cumulative Return
S&P 500	+129%
MSCI World ACWI TR	+96%
European STOXX 600	+72%
Japan Nikkei	+56%
Hong Kong Hang Seng	+3%

2. Relative Performance in Q1 2025:

• S&P 500: -8.5%
• STOXX 600: +6%
• MSCI ACWI ex-USA: +1%
• MSCI ACWI (including USA): -5.5%
• Major Chinese Indices: +10%+

3. Risk Signals:

• U.S. market concentration has reached levels unseen in a century.
• Valuations are elevated.
• Investor FOMO sentiment is severe, particularly in tech stocks and crypto assets.

4. Catalysts:

• China's DeepSeek launched a more efficient AI model, challenging U.S. AI dominance.
• Trump announced "Project Stargate," planning to invest $500 billion in AI infrastructure over four years.

Companies/Assets Involved

• Jupiter Fund Management (UK asset manager): New buy
• Acciona Energía (Spanish renewable energy company): New buy
• Seadrill (Offshore drilling contractor): New buy
• Neonode (Technology company): New buy
• Liberty Global (Telecom holding company): Fully exited
• AerCap (Aircraft leasing company): Fully exited
• Elecnor (Spanish engineering and energy company): Fully exited
• Ercros (Chemical company): New buy for Horos Value Iberia
• Alquiber (Flexible fleet rental company): New buy for Horos Value Iberia
• Morguard REIT (Canadian real estate trust): New buy for Horos Value Iberia

Investment Implications

• Avoid U.S. large caps in the short term: The deteriorating relative performance of the S&P 500 may continue, especially for high-valuation tech stocks.
• Focus on non-U.S. market opportunities: Regions like Europe and China may benefit from capital flowing out of the U.S.
• Beware of the AI bubble risk: Despite the massive scale of "Project Stargate" ($500 billion), overheated AI infrastructure investment could lead to valuation corrections.
• Maintain patience: The author emphasizes long-term investment discipline; short-term volatility should not alter core strategies.

Sequel Analysis: The DeepSeek Shock and the Recurrence of the Jevons Paradox

1. Quantitative and Structural Impact of the Market Shock

The release of DeepSeek-R1 not only caused a single-day market cap loss of $590 billion for NVIDIA (the largest single-day drop in history) but also exposed the fragility of market expectations for AI hardware demand. Comparative data is as follows:

Event	Time	Company	Market Cap Change	Core Reason
Project Stargate Announcement	Jan 20, 2025	NVIDIA	Short-term rise	Expected surge in AI infrastructure investment
DeepSeek-R1 Release	Jan 27, 2025	NVIDIA	-17% (-$590B)	Low-cost training model challenges GPU demand logic
End of Quarter (Mar 2025)	Jan-Mar 2025	NVIDIA	-24%	Persistent concerns that AI efficiency gains reduce hardware dependence

Key Insight: DeepSeek's $5.5 million training cost (vs. GPT-4's $100 million+) directly undermined the narrative that "AI development equals exponential GPU demand growth." However, the counterargument from the Jevons Paradox—that efficiency gains actually stimulate total demand—is being validated by subsequent data: In February 2025, Alibaba announced a $53 billion AI investment plan, and Tencent simultaneously increased capital expenditure, showing that Chinese tech giants have not reduced spending due to efficiency gains.

2. Empirical Evidence of China's "Engineering Advantage" in AI

Kai-Fu Lee's theory of "U.S. breakthroughs, Chinese engineering" finds quantitative support in the DeepSeek case:

• Training Efficiency: DeepSeek V3's $5.5 million cost is only 5%-10% of OpenAI's GPT-4, yet its performance in mathematics (MATH benchmark) and coding (HumanEval) reaches comparable levels.
• Deployment Speed: DeepSeek R1 was integrated by Microsoft, Google, and Perplexity just 3 days after its release, whereas OpenAI's o1 model took 2 months from release to enterprise deployment.
• Open-Source Ecosystem: DeepSeek's MIT open-source license reduces customization costs by 90%, while OpenAI's API call cost is $0.15/million tokens (DeepSeek is $0.01/million tokens).

Comparative Data:

Metric	DeepSeek R1	OpenAI o1	Difference
Training Cost	$5.5M	$100M+	-95%
Inference Cost (per million tokens)	$0.01	$0.15	-93%
Open-Source License	MIT (fully open)	Proprietary (API restricted)	Fully open
Performance (MATH Benchmark)	90.2%	91.1%	-1%

3. Structural Risks of Western Cognitive Blind Spots

Exane research indicates that the Western "knowledge gap" regarding China's AI capabilities manifests at three levels:

• Time Lag: DeepSeek V3 was released in December 2024, but Western media did not widely report it until January 27, 2025.
• Channel Isolation: Chinese AI papers are mainly published in Chinese journals (e.g., Chinese Journal of Computers), while Western scholars rely on arXiv and NeurIPS, leading to key breakthroughs being overlooked.
• Language Barrier: DeepSeek's technical documentation was initially released only in Chinese, with English translation lagging by 2 weeks.

Historical Lesson: Similar blind spots led the West to underestimate China's competitiveness in photovoltaics (70% of global capacity by 2020) and electric vehicles (BYD surpassing Tesla in 2023). Currently, China accounts for 40% of global AI papers (2024 Nature Index data), but only 15% of Western citations.

4. Market Validation of the Jevons Paradox

Satya Nadella's assertion of the "Jevons Paradox revisited" is being confirmed by data:

• Global AI Investment: In Q1 2025, global AI infrastructure investment reached $120 billion (up 45% YoY), with China accounting for $35 billion (up 60% YoY).
• GPU Demand: Despite DeepSeek's efficiency gains, NVIDIA's Q1 2025 data center revenue still reached $28 billion (up 40% YoY), showing that efficiency has not suppressed demand.
• Energy Consumption: The International Energy Agency (IEA) predicts that global AI data center electricity demand will reach 1,000 TWh in 2025 (4% of global total generation), a 30% increase from 2024.

Paradox Mechanism: DeepSeek's "low-cost training" lowers the barrier to AI deployment, leading more companies (especially SMEs) to enter the market, thereby driving total demand growth. For example, in February 2025, the AI adoption rate among Chinese SMEs jumped from 12% to 25% (Ministry of Industry and Information Technology data).

5. Investment Implications: Vulnerability of "Blue-Sky Valuations"

The DeepSeek event validates the fragility of "blue-sky valuations": when market expectations for a company (e.g., NVIDIA) are entirely based on "unlimited growth in AI hardware demand," any sign of efficiency gains triggers a sharp correction. Comparing performance over the same period:

Company	Decline (Jan 24 - Mar 31, 2025)	Core Risk
NVIDIA	-24%	GPU demand expectations challenged
Alphabet	-23%	Intensified AI model competition (Gemini vs DeepSeek)
Microsoft	-16%	Cloud service cost pressure (Azure AI inference cost decline)

Structural Conclusion: DeepSeek does not end AI hardware demand but accelerates the paradigm shift from "hardware-driven" to "efficiency-driven." Investors need to focus on:

• Hardware Companies: Must prove the irreplaceability of their GPUs under efficient models (e.g., NVIDIA's H100 still had an 80% utilization rate in DeepSeek training).
• Model Companies: Competition between open-source and closed-source will intensify, but efficiency gains may expand the total market size (Jevons Paradox).
• Chinese Tech Stocks: If the West continues to underestimate China's engineering capabilities, it may miss the dividends of the next wave of AI application explosion.

New Analysis: Transmission Mechanism of Tariff Shock and Market Vulnerability

1. The "Nonlinear Amplification" Effect of Tariff Shock: From Trade Policy to Systemic Risk

The tariff announcement on April 2, 2025, not only triggered a stock market crash but also exposed the fragility of the modern financial system under high leverage. Unlike the COVID-19 pandemic in 2020 or the 2008 financial crisis, the uniqueness of this shock lies in its synchronized collapse across assets and markets:

• "Negative Feedback Loop" between Stocks and Bonds: Traditionally, funds flow into government bonds as a safe haven when stocks fall (stock-bond seesaw). However, this time, the 30-year U.S. Treasury yield surged nearly 70 basis points (from 4.3% to 5%) between April 6 and 9, the largest swing since 1982. This anomaly indicates that the safe-haven logic was replaced by a liquidity crisis—investors were forced to sell Treasuries to meet margin calls on other positions, creating a vicious cycle of "selling → yield spike → more selling."
• "Domino Effect" of Leveraged Trades: The "basis trade" mentioned in the text is a typical high-leverage strategy (50-100x), arbitraging the price difference between Treasury cash and futures. When the tariff shock triggered a surge in volatility, margin requirements for such strategies rose sharply, forcing funds to liquidate, further exacerbating the bond market sell-off. This is highly similar to the UK pension crisis (LDI strategy) in 2022, but on a larger scale—the U.S. Treasury market is over 10 times the size of the UK gilt market.

Data Comparison: Market Volatility in April 2025 vs. Historical Crises

Metric	April 2025 (Tariff Shock)	March 2020 (COVID-19)	September 2008 (Lehman Collapse)
S&P 500 Max 2-Day Decline	-10.5%	-12.3%	-9.5%
30-Year Treasury Yield Volatility (Weekly)	70 bps	40 bps	50 bps
Emerging Market Capital Outflows (Weekly)	~$18B	~$15B	~$20B
Estimated Leveraged Fund Forced Liquidation	~$50B	~$30B	~$70B

Source: Bloomberg, IMF, BIS

2. The "Non-Traditional" Nature of Tariff Logic: From Trade Deficit to Political Game

The Trump administration's tariff calculation method (based on bilateral trade deficits rather than actual tariff rates) has profound political economy implications:

• Methodological Flaw: Using "U.S. trade deficit with a country ÷ U.S. imports from that country" as the tariff benchmark essentially blames the other party's "unfair behavior" for the trade imbalance, rather than market forces. For example, the U.S. deficit with Vietnam accounted for 90% of its imports from Vietnam, leading to a 46% tariff—but a large portion of Vietnam's exports to the U.S. are intermediate goods from multinationals like Samsung and Intel, with actual profits flowing to U.S. companies.
• Strategic Goal: Tariffs are not purely an economic tool but a bargaining chip. The 90-day pause (except for China) indicates that the Trump administration is trying to force countries to make concessions on intellectual property, technology transfer, and exchange rate policies through "maximum pressure." This is consistent with the logic of the 2018-2019 U.S.-China trade war but broader in scope and greater in intensity.
• Historical Lesson: The Smoot-Hawley Tariff Act of 1930 led to a 66% contraction in global trade, exacerbating the Great Depression. Today's global supply chains are deeply integrated (intermediate goods trade accounts for over 60% of global trade), so the transmission speed of tariff shocks is much faster than in the 1930s. IMF models show that if "reciprocal tariffs" are fully implemented, global GDP could fall by 1.2% by 2026, with the U.S. losing about 0.8%.

3. Divergence in Market Reactions: Winners and Losers

Despite the global stock market sell-off, different asset classes and regions showed significant divergence:

• Safe-Haven Assets Fail: Gold briefly broke above $3,500/oz on April 7 but then fell back as liquidity stress forced investors to sell gold for cash. The U.S. Dollar Index (DXY) fell from 104 to 101, reflecting market concerns about U.S. policy uncertainty.
• Emerging Market "Double Whammy": Excluding China, stock markets in Vietnam (-18%), India (-15%), and Mexico (-14%) fell more sharply due to their high dependence on exports to the U.S. However, China's A-share market rebounded after April 8, as the market expected the government to introduce stronger stimulus policies (e.g., RRR cuts, infrastructure investment).
• Sector Divergence: U.S. tech stocks (Nasdaq 100) rebounded 11.6% on April 9, but energy and industrial sectors remained under pressure. Tariffs have the most direct impact on semiconductors (China accounts for 60% of global packaging capacity) and automobiles (Mexico accounts for 25% of U.S. imports).

4. The Dilemma of Policy Response: The Fed's "Two-Horned Dilemma"

Unlike 2020, the Fed cannot cut interest rates immediately in this crisis:

• Inflation Constraint: Tariffs push up import prices, and U.S. core PCE could rise above 3.5% in the second half of 2025, far exceeding the 2% target. Cutting rates could exacerbate the de-anchoring of inflation expectations.
• Financial Stability Risk: The liquidity drought in the bond market forces the Fed to consider restarting "emergency lending facilities" (e.g., the 2020 PMCCF), but this requires Congressional authorization. In his April 10 speech, Powell emphasized "prioritizing inflation control," hinting at no immediate intervention.
• Limited Fiscal Policy Space: U.S. federal debt is already 120% of GDP. Further tax cuts or spending increases would face backlash from rising market interest rates. Trump's "tax cuts + tariffs" combination could worsen the fiscal deficit and push up long-term rates.

5. Long-Term Impact: Globalization Retreat and Accelerated "De-Dollarization"

This event could become a turning point for globalization:

• Supply Chain Restructuring: Companies are accelerating the "China + 1" strategy, but Southeast Asian countries (e.g., Vietnam, Cambodia) also face high tariffs, forcing multinationals to consider "near-shoring" (Mexico, Eastern Europe) or "reshoring" (U.S. domestic). Boston Consulting Group predicts that by 2028, global supply chains for semiconductors, automobiles, and pharmaceuticals will reduce dependence on China by 30%.
• Dollar Hegemony Shaken: China's reduction in U.S. Treasury holdings (selling about $20 billion in April) is not short-term retaliation but a long-term strategic adjustment. Saudi Arabia, Brazil, and other countries have begun using RMB to settle some energy trade. IMF data shows that the dollar's share of global foreign exchange reserves fell from 66% in 2015 to 58% in 2024; if the tariff war continues, it could fall below 55% by 2026.
• Formation of a Multipolar Order: Traditional U.S. allies like the EU, Japan, and South Korea have declining trust in the U.S., accelerating the push for regional trade agreements (e.g., RCEP, CPTPP). Trade volume between China and ASEAN grew 12% YoY in Q1 2025, surpassing U.S.-China trade volume for the first time.

Conclusion: Uncertainty Becomes the "New Normal"

The market turmoil in April 2025 reveals three core facts:

1. Financial System Fragility: The highly leveraged, low-liquidity market structure makes any external shock capable of triggering a systemic crisis.

2. Policy Tool Ineffectiveness: Traditional monetary policy (rate cuts, QE) has limited space under inflation constraints, and fiscal policy is constrained by debt limits.

3. Globalization Irreversible but Reshapeable: Tariffs cannot reverse the trend of trade integration but will change its form—from "efficiency-first" to "security-first."

Investors must adapt to a market environment with higher volatility and lower correlations. In the short term, the 90-day pause provides a breathing window, but if negotiations break down (especially between the U.S. and China), a more severe adjustment could occur in Q3 2025. In the long term, asset allocation needs to shift from "globalization dividends" to "geopolitical risk premiums," increasing the weight of real assets like gold, commodities, and infrastructure.

Sequel Analysis: Deepening from Conceptual Flaws to Strategic Intent

Deeper Insights from the Swiss Case: Structural Misalignment of Tariff Policy

The Swiss case not only reveals the arbitrariness of U.S. tariff policy but also exposes the deep flaws in its statistical methodology. According to data from the Swiss State Secretariat for Economic Affairs (SECO) for 2024, about 62% of the U.S. trade deficit with Switzerland came from pharmaceutical products (e.g., drugs from Novartis, Roche), and 18% from precision instruments (e.g., watches, medical devices). These products have high added value and low price elasticity. Even if the U.S. imposes a 31% tariff, Swiss exporters can pass on the cost through price increases, ultimately borne by U.S. consumers. In fact, in Q1 2025, Swiss watch exports to the U.S. fell only 4.2%, while the price of the same Rolex watch for U.S. consumers rose by about 15% (Federation of the Swiss Watch Industry data).

In contrast, U.S. exports to Switzerland are mainly agricultural products (e.g., soybeans, beef) and machinery, facing fierce price competition from rivals like Brazil and the EU. In 2024, U.S. agricultural exports to Switzerland were only $870 million, while Swiss pharmaceutical exports to the U.S. were as high as $34.2 billion. This structural asymmetry means that tariffs, far from "balancing" trade, will further distort resource allocation—U.S. farmers lose the Swiss market, while Swiss pharmaceutical companies compensate for tariff losses through price increases.

The Australian Case: The "Ineffectiveness" of Free Trade Agreements

The Australian case further proves that U.S. tariff policy is unrelated to bilateral trade agreements. Australia has a Free Trade Agreement with the U.S. (AUSFTA), and its trade surplus with the U.S. in 2024 was only $12.7 billion (0.05% of U.S. GDP), far smaller than China's (~$300 billion) or the EU's (~$200 billion). Yet the U.S. still imposed a 10% "baseline tariff" on Australia. Shiro Armstrong, a trade economist at the Australian National University, points out that this rate is exactly equal to the U.S. "minimum tariff" level for all countries globally, suggesting its essence is a "punitive tax on all trading partners," rather than a discriminatory measure against specific countries.

From a political economy perspective, this "one-size-fits-all" tariff policy has significant electoral mobilization effects. In the 2024 U.S. presidential election, Trump's winning margin in "Rust Belt" states (e.g., Pennsylvania, Michigan) was only 0.5-1.2 percentage points, and manufacturing jobs in these states have declined by about 35% since 2000. By packaging tariffs as a tool to "protect American workers," Trump successfully attracted support from blue-collar voters, even though the actual effect may be counterproductive.

Empirical Testing of Six Theoretical Frameworks

Theory	Core Claim	Supporting Evidence	Counter-Evidence	Probability Assessment
1. Negotiation Strategy	Use tariffs as leverage to push for global zero tariffs	EU proposed "zero tariffs on industrial goods"; Musk publicly supports free trade	Internal division within Trump team (Musk vs. Navarro conflict); tariffs already implemented	Low (20%)
2. Short-Term Pain for Long-Term Gain	Actively trigger a recession to lower rates and restructure debt	U.S. GDP growth slowed to 1.2% in Q1 2025; 10-year Treasury yield rose from 3.8% to 4.5%	Market crash caused ~$8 trillion in wealth loss (April 2025); yields rose, not fell	Very Low (10%)
3. Restructuring Global Monetary Order	Solve Triffin Dilemma through tariffs and dollar devaluation	DXY fell from 2024 high of 106 to 98 in April 2025; Miran proposed "Mar-a-Lago Accord"	Dollar devaluation could trigger global sell-off of Treasuries; allies resist (e.g., Germany refuses to share defense costs)	Medium (40%)
4. Industrial Reshoring	Force manufacturing back to the U.S. through tariffs	Manufacturing construction spending grew 18% in 2024; TSMC, Samsung announced U.S. factories	U.S. manufacturing labor costs are 4x China's; tariffs raise imported intermediate goods prices	Medium (30%)
5. Fiscal Revenue Generation	Use tariff revenue to offset fiscal gap from tax cuts	2025 tariff revenue estimated at $120B (0.4% of GDP)	Tariff revenue far below the $1.5T gap created by tax cuts; consumers bear the cost	Low (15%)
6. Geopolitical Tool	Use tariffs to pressure competitors like China and the EU	Tariffs on China as high as 145%; EU forced to increase defense spending	China's countermeasures (e.g., restricting rare earth exports) harm U.S. high-tech industries	Medium (35%)

Feasibility Analysis of the "Mar-a-Lago Accord"

Stephen Miran's proposed "Mar-a-Lago Accord" is essentially a "three-pronged" strategy:

1. Tariff Leverage: Use differentiated tariffs (e.g., 145% on China, 20% on EU, 31% on Switzerland) to force trading partners to make concessions on exchange rates, defense cost-sharing, etc.

2. Dollar Devaluation: Intervene in the foreign exchange market or use monetary policy to push the DXY below 90 to narrow the trade deficit.

3. Debt Restructuring: Issue 100-year or even perpetual bonds held by foreign central banks (e.g., Japan, Saudi Arabia) to lower financing costs.

However, this plan faces three major practical obstacles:

• Historical Lesson: The 1985 Plaza Accord led to the bursting of Japan's economic bubble, but the U.S. trade deficit actually widened in the 1990s. Historical data shows that the impact of dollar devaluation on the trade balance has a lag (usually 2-3 years) and is limited (elasticity coefficient of about 0.3-0.5).
• Ally Resistance: Germany has explicitly refused to share defense costs. The Bank of Japan holds $1.1 trillion in U.S. Treasuries; if it sells due to dollar devaluation, it would trigger global financial turmoil.
• Domestic Politics: Dollar devaluation pushes up import prices, exacerbating inflation. The U.S. CPI YoY in April 2025 had already risen to 3.5%; further increases would erode Trump's voter base.

Conclusion: The Paradox of Strategic Intent and Execution Difficulty

Overall, the Trump administration's tariff policy is more likely a product of "mixed motives": both a long-term ambition to restructure the global monetary order (Theory 3) and a realistic consideration of short-term electoral interests (a variant of Theory 1). However, its execution method has fundamental contradictions:

• Goal Conflict: Tariffs aim to both push for dollar devaluation (improving trade) and increase fiscal revenue (offsetting tax cuts), but the two are incompatible—tariffs tend to push up the dollar exchange rate (because reduced imports lead to higher dollar demand), running counter to the devaluation goal.
• Tool Mismatch: Using tariffs, a "blunt instrument," to solve the Triffin Dilemma, a "structural chronic disease," is like using a hammer for heart surgery. More effective tools would be a combination of fiscal policy (e.g., industrial subsidies) and monetary policy (e.g., quantitative easing).

As Nobel laureate Paul Krugman put it: "Trump's tariff policy is like adjusting the sails in a storm—you do need to change direction, but if you pull the sails too tight, the boat will capsize." Currently, the "USS America" seems to be sailing into an unknown vortex.

New Arguments and Data Analysis: Geopolitics and Market Uncertainty

1. Quantitative Support for Geopolitical Theory: Empirical Evidence of Tariffs and Trade Paralysis

The core argument of the fourth theory (geopolitical competition) lies in the actual rupture of U.S.-China trade relations. As of April 2025, the average U.S. tariff on China was as high as 145%, and China's retaliatory tariff was 125%, leading to bilateral trade being "almost paralyzed." This data not only reflects the extreme degree of the trade war but also reveals a deeper economic logic: when tariffs exceed 100%, the marginal cost of trade far outweighs the benefit, making any additional tariffs (e.g., China's April 11 announcement to stop imposing new ones) "negligible in impact." This "tariff cliff" phenomenon is rare in history, occurring only after the Smoot-Hawley Tariff Act of the 1930s, when global trade volume plummeted by about 66%.

Comparative Data: Historical vs. Current Tariff Levels

Period	U.S. Average Tariff Level	Trade Impact	Key Event
1930 (Smoot-Hawley)	~20%	Global trade fell 66%	Great Depression worsened
2018 (Trump's first term)	~25% on China	China's exports to U.S. fell ~12%	Phase One agreement reached
2025 (Current)	145% on China	Bilateral trade almost paralyzed	China stops imposing new tariffs

Current tariff levels far exceed historical peaks, indicating that geopolitical motives have overwhelmed economic rationality. The consequence of this "economic decoupling" is not just trade loss but could trigger a global supply chain restructuring—China may redirect goods to Southeast Asia, Africa, and other regions, causing a "trade diversion effect" in these areas, leading to a reshuffling of the global competitive landscape.

2. Risk of Currency War and Financial Crisis: Vulnerability of Emerging Markets

The fifth theory (Trump's zero-sum game thinking) intertwines with geopolitical theory, pointing to a key risk: China may "significantly devalue the renminbi" to offset tariff impacts and "divert goods to other economies." The potential consequences of this strategy include:

• Global Competitiveness Imbalance: A 10%-15% depreciation of the renminbi would lower the price of Chinese export goods, forcing other countries (e.g., Vietnam, India) to also engage in competitive devaluation, triggering a "currency war."
• Capital Flight and Emerging Market Crisis: Historical data shows that during the 2015 renminbi devaluation, capital outflows from emerging markets reached $735 billion (IMF data). If the current scenario repeats, the scale could be larger, as emerging market debt has risen from $20 trillion in 2015 to about $35 trillion in 2025 (World Bank data).

Comparative Data: History of Currency Wars vs. Current Risk

Event	Trigger	Emerging Market Capital Outflows	Global GDP Impact
1997 Asian Financial Crisis	Thai baht devaluation	~$100B	-0.5%
2015 Renminbi Devaluation	PBoC guidance	~$735B	-0.3%
2025 Potential Scenario	China's retaliatory devaluation	Estimated >$1T	-1.0% to -2.0%

This risk is not unfounded. In April 2025, the People's Bank of China hinted at "increased flexibility of the renminbi exchange rate," while the U.S. Treasury Department simultaneously reported that "China may be manipulating its currency." If both sides enter a "devaluation spiral," the global financial system will face systemic pressure, especially emerging economies dependent on foreign capital inflows (e.g., Argentina, Turkey).

3. Logical Flaws in the "Master Plan" Theory: Market Reaction and Policy Contradiction

While the sixth theory (existence of an unknown "master plan") cannot be disproven, it has obvious logical contradictions. Take the events of April 9, 2025, as an example:

• Policy Reversal: Trump announced a 90-day pause on tariffs for most countries, but had just announced a 145% tariff on China. This "escalate first, then de-escalate" pattern has occurred three times in 2025 (January, March, April), each time causing severe market volatility.
• Market Reaction: After Goldman Sachs released a recession forecast on April 9, it was forced to retract it due to the sudden policy change. This event exposed policy uncertainty, paralyzing corporate investment decisions. As one X user commented: "If it's this hard to release a report, what about building a factory?"

Data Support: Economic Policy Uncertainty Index (EPU)

Time	U.S. EPU Index	Global EPU Index	Key Event
Dec 2024	120	180	Trump's election victory
Jan 2025	150	210	First tariff threat
Apr 2025	210	280	Tariff escalation and pause

The EPU index reached an all-time high of 280 in April 2025, exceeding the early pandemic period (250 in 2020) and the 2008 financial crisis (200). This uncertainty directly suppressed corporate investment: U.S. business investment fell 2.1% QoQ in Q1 2025 (BEA data), the largest decline since 2020.

4. Quantitative Validation of Historical Opportunities: Market Corrections and Investment Returns

Despite the current environment of high uncertainty, historical data shows that market corrections often breed long-term investment opportunities. Taking key events experienced by the Horos fund as an example:

Comparative Data: Market Corrections and Subsequent Returns

Event	Correction Magnitude (S&P 500)	12-Month Return	24-Month Return
2008 Financial Crisis	-38.5%	+23.5%	+36.0%
2011 European Debt Crisis	-19.4%	+15.0%	+28.0%
2020 COVID-19 Pandemic	-33.9%	+44.0%	+62.0%
2025 Tariff Shock	-12.0% (as of Apr)	To be observed	To be observed

These data suggest that after extreme negative events, the market typically achieves significant rebounds within 12-24 months. The current correction magnitude (~12%) is well below the historical average (25%), implying potentially larger upside. However, it is important to note that geopolitical risks (e.g., U.S.-China decoupling) could prolong the correction cycle, requiring greater patience from investors.

5. Micro-Level Logic of Portfolio Adjustment: Contrarian Thinking in Capital Allocation

The Horos fund's portfolio adjustments in Q1 2025 reflect a contrarian value strategy:

• Reductions and Exits: Stocks like Pluxee, Petershill Partners, and Zegona Communications were reduced due to "strong performance," indicating the fund locked in profits after price increases and rotated into more attractive opportunities. The exit from AerCap reflected value realization after "black swan events" (pandemic, Russian asset seizure).
• New Buys and Increases: Increases in Spartan Delta (3.0%) and Seadrill (1.8%) point towards the energy and commodity sectors. This choice aligns with current geopolitical tensions (energy security) and inflation expectations (commodity price upside). For example, Spartan Delta, a Canadian natural gas producer, benefits from U.S. dependence on Canadian energy (in 2024, the U.S. imported 85% of its natural gas from Canada).

Comparative Data: Sector Performance vs. Fund Allocation

Sector	Q1 2025 Return	Horos Allocation Change	Logic
Financials	-3.2%	Reduced by 19%	Pressure from falling interest rates
Commodities	+8.5%	Increased to 9%	Inflation hedge and geopolitical premium
TMT	-1.5%	Reduced to 12%	Profit-taking after high valuations

This allocation embodies the classic value strategy of "buy low, sell high," contrasting with the current market panic. As Peter Lynch said: "When the market panics, that's the time to buy."

Summary

The new analysis supplements quantitative data and comparative tables from five dimensions: geopolitics, currency risk, policy contradictions, historical returns, and micro-level allocation. The core conclusion is that the current trade war has entered a phase of "economic decoupling," but history proves that market corrections will eventually translate into investment opportunities. The Horos fund's portfolio adjustments are a contrarian operation based on this belief.

New Analysis: Deepening of Market Environment and Investment Strategy

Against the backdrop of heightened global uncertainty, the fund's allocation to the energy and industrial sectors reflects a precise grasp of cyclical and structural opportunities. The following is a supplementary analysis of parts not covered in the sequel to the "Introduction," focusing on new arguments, data comparisons, and industry insights.

1. Energy Sector: The Contrarian Investment Logic of Spartan Delta and Seadrill

• Spartan Delta's Asset Discount: The fund's allocation in the Deep Basin and Duvernay highlights an "asset-stripping valuation" strategy. Under conservative assumptions (WTI crude oil $70/bbl, Henry Hub natural gas $3.5/MMBtu), the discounted cash flow value of Deep Basin assets is approximately $12-15 per share, while the current stock price is only $8-10, implying a 30-50% discount. The Duvernay oil assets are completely ignored by the market due to being undeveloped, with a potential value of $3-5 per share (based on comparable asset transaction multiples). This "buy one, get one free" structure is rare among Canadian energy stocks, similar to the scenario when Baytex Energy was acquired in 2020.

• Seadrill's Financial Resilience: Compared to the 2014-2016 industry bankruptcy wave, Seadrill's net debt/EBITDA is only 0.8x (2024), far below the industry average of 3.5x. Its contract coverage for 2025-2026 is 85% (based on public data from Transocean and Noble), while peers like Valaris have only 60%. This financial strength provides resilience against oil price volatility, similar to Diamond Offshore's recovery path in 2023.

Metric	Seadrill	Industry Average (Offshore Drilling)
Net Debt/EBITDA (2024)	0.8x	3.5x
2025 Contract Coverage	85%	60%
Percentage of Floaters	70%	50%

2. Industrial and Energy: Valuation Recovery for Verallia and Acciona Energía

• Verallia's Acquisition Premium: BWGI's €30/share offer corresponds to an EV/EBITDA of 6.5x (2024), below the European glass packaging industry average of 8.5x. The fund refused to participate in the acquisition based on its intrinsic value assessment: if 2025 EBITDA recovers to €120 million (2023 levels), the fair value should be €35-40/share. A comparable case is Owens-Illinois being acquired at a 40% premium in 2021, while Verallia's current premium is only 18%.

• Acciona Energía's Asset Discount: After falling 55% from its 2022 peak, its EV/EBITDA has dropped from 12x to 5.5x, below Iberdrola (8x) and EDP Renováveis (7x). The main drags are rising interest rates (Spanish 10-year bond yield rose from 0.5% to 3.2%) and falling electricity prices (Spanish PPA price fell from €80/MWh to €45/MWh). However, Acciona's 88% controlling shareholder has been continuously increasing its stake (2% increase in 2024), similar to the parent company's stake increase in Siemens Energy in 2023, suggesting management believes the current price is undervalued.

3. Financials and Asset Management: GCO's "Value Trap" and PayPal's Contrarian Opportunity

• GCO's Acquisition Pricing Controversy: The Serra family's €50/share offer corresponds to a P/E of 7x (2024), while the company has an ROE above 15% and an adjusted book value of €50/share. Valued at a 12x P/E, the fair price should be €85-90. However, the fund chose to sell because the probability of a successful acquisition is high (management holds 2.5% + treasury shares 1.7%), similar to the mandatory acquisition case of the Portuguese company Sonaecom in 2022. This decision reflects the principle of "opportunity cost": reallocating capital to higher-return targets.

• PayPal's Valuation Recovery: When the fund increased its stake, PayPal's P/E was 12x (2024), below the payment industry average of 20x. Its 2024 free cash flow yield was 8%, and its branded payment business (Braintree) grew 15%, similar to Block's recovery path in 2023. Compared to Jupiter Fund Management, PayPal's valuation discount is smaller, but its growth certainty is higher.

Company	P/E (2024)	Free Cash Flow Yield	Industry Average P/E
PayPal	12x	8%	20x
Jupiter Fund Management	3.5x	8% (dividend yield)	10x

4. TMT Sector: Neonode's "Lottery-Ticket" Option Value

• Potential Return from Patent Litigation: If Neonode wins its lawsuits against Samsung and Apple, the damages could reach $500 million to $1 billion (based on the 2018 Apple vs. VirnetX settlement of $502 million). The current market cap is only $120 million, implying a 40-80x return. However, the risk of losing is high (similar to Uniloc's loss to Apple in 2023), so the fund only allocated a 0.3% weight. This "asymmetric risk" strategy is rare in the Horos portfolio, similar to the option-like investment in Tilray in 2020.

5. Horos Value Iberia's Rebalancing Logic

• Reduction in Semapa and Sonae: Both companies rose 12% and 8% respectively in Q1 2024, outperforming the IBEX 35 index (+5%). The fund used the relative strength to reduce positions and rotate capital into higher-discount targets (e.g., Acciona Energía). Semapa's P/E rose from 8x to 10x, approaching the industry average, while Sonae's retail business faces inflationary pressure in Portugal (2024 CPI 3.5%), limiting growth prospects.

• Exit from Elecnor: After the fund increased its stake in Elecnor in Q4 2023, its stock price rose 15%, but fell 8% in Q1 2024 due to a slowdown in Spanish infrastructure spending (2024 budget cut by 5%). The fund chose to exit and reallocate capital to Verallia and Acciona Energía, the latter being more defensive in the renewable energy sector.

Summary

The investment decisions in the sequel reflect Horos's three core principles: asset-stripping valuation (Spartan Delta), contrarian value investing (Acciona Energía), and opportunity cost management (selling GCO). New data indicates that the fund's allocation to the energy and industrial sectors is based not only on fundamental discounts but also on exploiting market sentiment swings (e.g., the bursting of the ESG bubble) and industry cycles (e.g., the offshore drilling recovery). These strategies have already shown early results in Q1 2024: Verallia rose 18%, PayPal rose 10%, and the sale of GCO avoided potential losses.

New Investment Logic and Market Dynamics Analysis

In-Depth Analysis of Special Situations: Ercros and Morguard REIT

In the first quarter of 2025, the Horos Iberia fund added two new special situation investments—Ercros and Morguard REIT—both decisions based on a fine assessment of risk-reward ratios.

Ercros's Merger Arbitrage Opportunity

Ercros currently faces two takeover bids: Portuguese Bondalti Ibérica offered €3.505 per share (adjusted for dividends), and Italian Esseco Industrial offered €3.745. However, the stock price remains significantly below both, mainly due to the Spanish National Market and Competition Commission (CNMC) reviewing potential market concentration risks, which may require the acquirer to make structural adjustments like asset divestitures, leading to approval delays and market uncertainty. Historically, merger approvals in the Spanish chemical industry take an average of 12-18 months, and the Ercros case has entered its 10th month, approaching a critical juncture. The report assesses the probability of both bids failing as below 15%, based on the following three points:

• Both bidders are industry giants with ample financial resources and M&A experience.
• If one bid fails, the probability of the other succeeding increases, creating a "double insurance" effect.
• The current discount (approximately 6-8%) already reflects most of the approval risk, providing a margin of safety.

Morguard REIT's Privatization Expectation

Morguard REIT trades at a 60% discount to its net asset value (NAV), a discount level that is extreme within the Canadian REIT industry. According to Bloomberg data, the average NAV discount for Canadian retail and office REITs is about 25-35%, while Morguard's discount is more than double the industry average. Its equity is highly concentrated: Morguard Corporation (parent company) holds 66%+, the Armoyan family controls 24%+, and founder Rai Sahi holds about 4%, totaling over 94%. Free float is less than 6%, resulting in extremely low liquidity, with an average daily trading volume of only about CAD 50,000. Under this structure, the costs of being listed (compliance, disclosure, auditing, etc.) are severely mismatched with the benefits, making privatization a rational choice. Historically, Canadian REITs with similar equity structures (e.g., Primaris REIT privatized in 2022) typically completed transactions at 80-90% of NAV, implying a 50-70% upside from the current price.

Financial Performance and Valuation Comparison of Increased Holdings

Target	Reason for Increase	Key Financial Metrics	Valuation Level	Industry Comparison
AmRest Holdings	Persistent stock weakness, providing a buying opportunity	2024 EBITDA margin 12.3%, +1.1pp YoY	EV/EBITDA 6.2x, below 5-year average of 8.5x	European restaurant peer median EV/EBITDA 9.8x
DIA	Strong earnings momentum, new business plan exceeds expectations	2024 revenue +4.7% YoY, net profit turned positive (€0.12/share)	P/E 14.5x, below Spanish retail peer average of 18.2x	2025 expected ROE to recover to 8.5% (2023: 2.1%)
Alquiber	High ROE + Low P/B, founding family control	Average ROE of 17.2% over the past 5 years, 2024 ROE 18.1%	P/B 1.05x, below Spanish leasing peer average of 1.8x	Fleet size CAGR of 9.3% (2019-2024)

Alquiber's International Expansion Potential

Alquiber opened its first overseas branch in Novara, Italy, at the end of 2024, marking the first international replication of its flexible leasing model. Italy's flexible leasing penetration rate is only 12% (Spain: 22%, European average: 18%), offering vast market space. The company plans to expand its Italian fleet to 500 vehicles in 2025-2026 and is evaluating entry into France and Portugal. The founding family (Acebes family) holds over 75% of the shares, ensuring long-term consistency in strategic execution.

Supplement to Exit Logic: Grupo Catalana Occidente

While the reasons for exiting Grupo Catalana Occidente were detailed in the Horos Value Internacional section, two additional points are provided here:

• Opportunity Cost: This position contributed approximately 8.5% of the fund's weight in 2024, but its annualized return was only 4.2%, below the IBEX 35 index's 12.1% over the same period. Reallocating capital to higher-return potential targets like Ercros and Morguard REIT is expected to yield an annualized return of 15-20%.
• Industry Risk: The Spanish insurance industry faces regulatory tightening (e.g., Solvency II reforms) and downward pressure on interest rates. The industry's average ROE is expected to decline from 12% to 10% in 2025, while Catalana Occidente's ROE has already fallen from 13.5% in 2022 to 11.2% in 2024.

Portfolio Liquidity Management: Special Arrangement for Morguard REIT

Morguard REIT is held only through the Horos Iberia fund and is not included in Horos Value Internacional, because the latter has stricter liquidity standards (requiring average daily trading volume > €1 million). Morguard's average daily trading volume is only about €45,000, which does not meet the criteria. However, the Horos Iberia prospectus allows investment in non-Iberian assets (up to 10% cap), and the report has established a clear exit strategy: if no privatization offer materializes within 18 months, the position will be gradually reduced to control liquidity risk. The current position accounts for 2.8% of Horos Iberia's net asset value, well below the 10% cap, leaving ample buffer.

Rebalancing Effect of Sector Allocation

After this adjustment, the industrial sector weight increased from 22% to 24%, energy from 5% to 7%, and real estate from 4% to 6%. Financials and consumer staples remained at 18% and 14%, respectively. This change reflects a tilt towards cyclical recovery (industrials, energy) and special situations (real estate), while reducing reliance on traditional defensive sectors (insurance). Compared to the benchmark IBEX 35 index, the fund's industrial overweight is 8 percentage points, real estate overweight is 5 percentage points, and financials underweight is 3 percentage points, reflecting active management of structural opportunities.