2nd Quarter 2025 Commentary

Theme & Background

This chapter fundamentally questions index investing, a core strategy in modern markets. The author argues that since the inception of ETFs (approximately 25 years ago), their actual returns have fallen far short of the 10% annualized return benchmark presupposed by the Capital Asset Pricing Model (CAPM). This reality poses a severe challenge to the theoretical foundation of asset allocation.

Core Thesis

The author's central argument is that index investing has not only failed to beat active management but has also failed to achieve its own presupposed 10% annualized return target. A counter-intuitive, contrarian market judgment is: Over the past two decades, passively holding a single physical asset (gold) has systematically outperformed all regional stock index ETFs. The report criticizes index investing for its inherent flaw of "anti-scarcity," meaning that ETFs continuously create new shares to absorb new capital, theoretically denying the value of overvaluation and asset scarcity.

Key Arguments & Data

The author cites a series of ETF return data from their inception up to May 2025, comparing them with gold and Bitcoin.

Annualized Return Comparison of Various Asset ETFs Since Inception:

Asset Class / ETF	Return Period	Annualized Return	Core Conclusion
iShares Core S&P 500 ETF	25 Years	~8%	Below the expected 10% benchmark
MSCI EAFE ETF (Developed Markets)	24 Years	Below 6%	Far below expectations
iShares Core MSCI EM ETF (Emerging Markets)	Since Inception	8.8%	Did not approach the 10% target
Regional Indices (Europe, Asia, Latin America)	20+ Years	None approached 10%	Systematic failure
Bond ETFs	20+ Years	1.9% - 3.7%	Negative returns after tax and inflation adjustment
iShares Gold Trust	20 Years	10.4%	Outperformed all stock indices
Gold Spot (Since May 2000)	25 Years	10.4%	Achieved the index's presupposed target

Other Key Data & Logic:

• Disappearance of Profit Support Factors: The report points out that factors supporting high corporate profit margins and valuations in the past (low interest rates, labor cost arbitrage, low taxes, falling commodity prices) are receding. Future index allocations will lose these environmental tailwinds.
• Scarcity vs. Anti-Scarcity: The operational mechanism of ETFs is "anti-scarcity"—unlimited issuance of new shares for incoming capital, which destroys the scarcity value of the underlying assets. In contrast, Bitcoin is based on explicit scarcity (a total supply of 21 million coins).
• USD vs. CHF: Since 1971, the US dollar has depreciated over 80% against the Swiss franc. The root cause lies in different money supply growth rates: US M2 money supply grew 30-fold (6.7% annualized), while Switzerland's grew only 10-fold (4.6% annualized). This illustrates the value preservation capability of a scarce (slower supply growth) currency.
• Bitcoin's Scale: As of July 30, 2025, Bitcoin's market cap reached $2.3 trillion (equivalent to the 8th largest component of the S&P 500). All Bitcoin ETFs combined total $160 billion. If merged into a single fund, it would be the eighth-largest US stock ETF.

Companies/Assets Involved

• iShares (BlackRock): Multiple ETF samples in the report (e.g., IVV, EFA, EEM, GLD) are from iShares, used as negative examples of index investing.
• Gold: Represented as a successful scarce asset, its iShares Gold Trust (GLD) achieved a 10.4% annualized return.
• Bitcoin: Presented as a modern example of scarcity value investing, with a massive market cap and significant ETF fund size.
• MSCI EAFE / MSCI Japan ETF: Used to critique the geographic mismatch problem of index investing. The author argues that the iShares MSCI Japan ETF (EWJ) is not a true exposure to the Japanese domestic economy; its components are mostly multinational corporations, with actual exposure to major export markets (e.g., the US).

Investment Implications

For investors, the report advocates for a thorough re-evaluation of passive index allocation strategies.

1. Beware of the "Index Illusion": Do not presuppose a 10% annualized long-term return for index ETFs. Actual data shows stock index returns over the past 20 years have been far lower, and bond returns have been negative after tax and inflation adjustments.

2. Embrace "Scarcity Investing": Investors should actively seek and allocate to assets with scarcity value (e.g., specific physical resources, cryptocurrencies, or other supply-constrained targets) rather than passively buying market-cap-weighted indices that continuously dilute shares.

3. Question Consensus Exposure: Investors should realize that buying a regional index ETF like MSCI Japan does not provide exposure to the local economy but rather a portfolio of global major export markets. This suggests investors need more granular analysis of the underlying assets' actual economic exposure.

The Hidden Value of the Japanese Market: Valuation Distortion from Non-Consolidated Reporting

John Templeton's core insight when investing in Japanese stocks during the 1950s-1960s was discovering a systematically overlooked accounting standard difference. At the time, Japanese companies were not required to consolidate subsidiary financial statements, making their reported P/E ratios appear normal. However, when adjusted for US standards (including subsidiaries' proportional earnings), the actual P/E was only "a few times." This structural arbitrage opportunity is something index funds cannot capture—indices rely on public data, while adjusted true valuations require deep company analysis.

Templeton Growth Fund vs. S&P 500 Ten-Year Return Comparison (1970-1979)

Metric	Templeton Growth Fund	S&P 500
Cumulative Total Return	~3x	1x (Benchmark)
Annualized Return	18.8%	6.5%

Data Source: etfdb.com (Original Footnote 4)

Notably, Templeton consistently underperformed the S&P 500 in the prior decade (1955-1968) but maintained a ~50% allocation to Japan, which only paid off in the 1970s. This reveals the core challenge of active management: the time dimension and career risk. If managed by external investors, the fund might have been withdrawn due to poor short-term performance.

Current New Valuation Anomalies in Japan: Untouchable by ETF Crowds

Unlike the Templeton era, Japan's current structural distortions stem from market reforms after 30 years of inefficiency. Horizon Kinetics identifies two unique strategic opportunities in Japan:

1. Unsustainably Deep Discounts: Some Japanese companies, due to market overpricing of governance deficiencies, trade below book value (P/B < 0.5) with ample cash flow. As the Tokyo Stock Exchange pushes for capital efficiency reforms (e.g., requiring companies with P/B < 1 to propose improvement plans), these discounts have mean-reversion potential.

2. Different Value Realization Mechanisms: Japanese companies tend to release value through increased dividends and share buybacks rather than US-style M&A and spin-offs. For example, some traditional manufacturers hold significant hidden assets (land, cross-shareholdings) not fully reflected in accounting standards.

"Reverse Active vs. Passive" Argument: Index's Structural Disadvantage is Irreversible

The original text presents a key reversal: based on the US economy shifting from a profit acceleration phase (past 20 years) to structural slowdown, and the distortion of market structure by indexation, indices can no longer outperform active management. New supporting arguments are as follows:

Index's "Overcrowding" in Weighted Sectors and "Marginalization" of Scarce Sectors

Using S&P 500 sector weight changes as an example:

Sector	1990 Weight	2025 (Approx.)	Weight Change
Information Technology (incl. Amazon/Meta/Alphabet)	~6%	44%+	+38pct
Utilities	6.2%	2.3%	-3.9pct
Basic Materials (Copper, Iron, Gold, etc.)	~3%	<0.5%	-2.5pct+
Gold & Silver Companies (Newmont+Freeport)	Not separately tracked	0.24%	Near zero

Data Source: Original text and Horizon Kinetics research

This "crowding-out effect" means that if utilities or basic materials become growth sectors driven by AI data center expansion, the index will barely capture their excess returns. For instance, NRG Energy's stock price rose 4x since ChatGPT's launch (mentioned in the original text), but utilities' index weight has dropped from 6.2% to 2.3%, making future gains negligible for the index.

Scarce Resources: Supply-Side Constraints "Invisible" to Indices

The original text highlights "limiting factors" not represented by indices: water, land, natural gas—physical bottlenecks for data center expansion. For example:

• Water Resources: A single data center facility can consume up to 1 million gallons of water daily. Water rights in the western US are complex and opaque, and related companies (e.g., American Water Works) are not core S&P 500 holdings.
• Land & Natural Gas: Data center electricity demand relies on natural gas peaking plants, but pipeline and storage facility companies (e.g., Williams Companies) have extremely low index weights (~0.1%).
• Copper & Silver: As critical materials for electricity and connectivity, production capacity has not increased in a decade, yet the S&P 500 has only two representative stocks with a combined weight of 0.24%.

Stock Exchanges in Active Management: A Reflexive Case

The original text argues that holding stock exchanges (e.g., CME, LSE, HKEx) can outperform local indices over the long term. New data supplement:

Exchange vs. Corresponding Index Long-Term Return Comparison

Asset	Time Horizon	Cumulative Return Multiple (Exchange/Index)	Notes
CME Group vs S&P 500	Past ~20 Years	Outperformed S&P 500	Specific multiple not disclosed, but original text states "better off"
London Stock Exchange vs FTSE 100	Past ~20 Years	3x	Predatory pricing power, market data revenue
HKEx vs Hang Seng	Past 20 Years	3x	China IPO monopoly, Stock Connect mechanism

Logic: Exchanges' diversified revenue streams (trading fees, clearing, data, listing fees) make them akin to a "total economic index," with natural monopoly and pricing power. Traditional index funds can only passively hold constituent stocks, not own the trading infrastructure itself.

Latest Prediction for the Future of Indexation: Maximizing Opportunity Cost

Based on the above analysis, a quantitative framework can be distilled: the "opportunity cost" of indices is expanding in the following areas:

Dimension	Index Performance	Potential Increment from Active Management
Limiting assets not represented by indices (water, land, natural gas)	0% weight	May outperform index stocks over the next 10 years
Growth sectors with severely compressed weights (utilities, metals)	Weight < 2.5%	If growth exceeds expectations, index gets only marginal contribution
Structural valuation distortions (Japan, emerging markets)	Cannot identify accounting differences	Potential multi-fold returns over the long term (e.g., Templeton case)
Innovative revenue models (ownership-based royalty fees, e.g., Jerick Bernstein)	No corresponding product	Starting yield of 8%+, far exceeding index dividend yield

Conclusion: The "free ride" effect of indexation has disappeared after the mass adoption of ETFs—current market pricing is dominated by passive capital, which ironically creates pricing biases that active investors can systematically exploit.

New Arguments: The Dual Robot Dilemma of Indexation and AI

1. Structural Deficiencies of Indexation and AI: Lack of True Intelligence

• Limitations of Robot Models: Indexation and AI (e.g., large language models) are essentially "robot models" whose performance depends entirely on the quality and scope of input data. They cannot achieve true intelligence (e.g., Artificial General Intelligence, AGI) and therefore cannot identify unstructured, non-linear opportunities in stock selection like the best human investors.
• Ironic Risk: The biggest threat currently pushing the indexation system into dangerous territory is precisely the emerging AI companies themselves. These companies are core investment directions for the largest IT firms in the S&P 500 (e.g., Microsoft, Google, Amazon), but their AI capabilities could conversely "disintermediate" the business models of these same IT giants. For example, generative AI could replace search engine advertising revenue models (Google), cloud service middle layers (Amazon AWS), or software subscription services (Microsoft Office).

2. Extremely Low Index Exposure to "Hard Assets": Data Comparison

• Natural Gas Exposure: The only direct natural gas hard asset exposure in the S&P 500 (via Texas Pacific Land Corp, TPL) is 0.009%. Even including asset-intensive companies like ExxonMobil and Chevron, total exposure is only slightly below 0.60%. This is far below energy's actual share of the overall economy (US natural gas accounts for ~32% of primary energy consumption).
• Water and Land Exposure: Direct water exposure is nearly zero (except TPL), and land exposure is similar. This absence means the index cannot capture profit growth from rising commodity prices, which simultaneously erodes the operating margins of cloud and AI companies (e.g., rising data center energy costs).

Asset Class	Direct Exposure in S&P 500 (Estimate)	Actual Share in US Economy (Reference)
Natural Gas (Hard Asset)	0.009% (TPL only) to 0.60% (incl. Exxon/Chevron)	~32% (Primary Energy Consumption)
Water (Direct)	Near zero (TPL only)	~0.5% (Water treatment in Utilities)
Land (Direct)	Near zero (TPL only)	~10% (Real Estate + Agriculture)

3. "Disintermediation" Risk Outside the Index: Blockchain and Stablecoins

• WisdomTree's Prime Digital Wallet: This platform uses tokenization technology, allowing users to bypass banks and payment processors (e.g., Visa) for spending, saving, and investing. This model directly threatens the financial and payment sectors within the S&P 500 (e.g., JPMorgan Chase, Visa, Mastercard).
• Disruptive Nature of Stablecoins: Stablecoins require only a collateral custodian, without a dominant card processor or bank "rails." By eliminating barriers to entry, they can commoditize banking and payment services, thereby reducing the profit margins of existing financial giants.
• Index Gap: These blockchain development companies (e.g., WisdomTree) are not included in the S&P 500, yet they represent a "disintermediation, substitution, or obsolescence risk" to core index components (financials, technology).

4. Indexation is No Longer a Total Economic Solution

• Historical Role: The S&P 500 was originally intended as a fully inclusive, reasonably proportional representation of the economy (though limited by the availability of publicly listed stocks, with real estate being a major exception).
• Current Deficiencies: Today, the index almost entirely excludes truly important sectors, including:
• Mature Industries: Extremely low exposure to hard assets (energy, water, land).
• Emerging Growth Industries: Disruptive technology companies like blockchain and decentralized finance (DeFi) are not included.
• Trapped at the Top: Although not superficially obvious, the top of the index (dominated by IT giants) is under dual siege from global competition outside the index and internal AI disintermediation. Active management can exploit these opportunities outside the index, while indexation cannot.

5. Data Sources & Supplementary Notes

• Natural Gas Exposure Calculation: Based on Statista data, TPL's weight in the S&P 500 is 0.03%, and approximately 30% of its third-party oil and gas production comes from natural gas, resulting in a direct exposure of 0.009%. Including ExxonMobil and Chevron brings the total weight to 1.92%, with natural gas accounting for over 30% of their weighted average production, resulting in total exposure slightly below 0.60%.
• Water and Land Exposure: Apart from TPL, there are almost no direct water or land asset companies in the S&P 500 (e.g., water utility companies or Real Estate Investment Trusts REITs are typically classified under other sectors).

Conclusion

The indexation system is facing a dual robot dilemma: it can neither capture the profit growth of hard asset companies (due to rising commodity prices) nor defend against the disruptive risks posed by AI and blockchain to its core components. The S&P 500 is no longer a total economic solution but a passive tool dominated by IT giants and exposed to structural deficiencies. Active management can mitigate these risks by leveraging global opportunities outside the index (e.g., hard assets, blockchain, emerging markets).

---

Theme and Background

This chapter explores the competitive threats faced by the Magnificent 7 technology giants, particularly from emerging "Private Mag 7" companies. The report argues that the current extreme market concentration in the Mag 7 (representing 33.2% of the S&P 500 index weight) overlooks the risk of fundamental erosion to their business models.

Core Thesis

The author's core judgment is: The Private Mag 7 (e.g., OpenAI, SpaceX) pose a substantive competitive threat to the Public Mag 7, not merely a "potential risk" in market narratives. The counterintuitive aspects are:

• The market generally believes the Mag 7's "moats" are unassailable, but the author points out their high profits stem from a prolonged lack of competition and an asset-light model, a foundation that is now crumbling.
• The "leader forever" assumption implicit in index investing could be overturned, as new entrants attack core profit models (e.g., Google's ad keyword revenue).

Key Arguments and Data

1. Market Concentration and Historical Comparison

• Mag 7 total market cap: $21.3 trillion, representing 33.2% of the S&P 500, unprecedented historically.
• Visa and MasterCard (the 13th and 16th largest S&P 500 weights) together account for 2.0%, also considered "dominant" companies.

2. Specific Paths of Competitive Threat

• OpenAI vs Alphabet: ChatGPT's free, ad-free model directly threatens Google's ad keyword auction business. Google's profits are highly dependent on "non-trivial searches" (high-value keywords), while ChatGPT may siphon off the most profitable traffic while leaving low-value trivial searches (which generate no revenue). Sora's video generator threatens YouTube's content advertising model.
• Rising Data Costs: Scale AI's subscription fee (averaging $93,000 per year) shows that the cost of data acquisition in the AI era has shifted from "zero cost" to a paid model. Meta's $30 million penalty for copyright infringement highlights increasing legal risks.
• Asset Model Shift: Data center projects like Stargate ($500 billion) and Stargate UAE force the Mag 7 to transition from "asset-light, high ROIC" to "asset-heavy, low free cash flow" models. OpenAI's cloud storage needs could directly compete with Amazon Web Services and Microsoft Azure.

3. Key Data Comparison

Metric	Public Mag 7	Private Mag 7
Total Market Cap	$21.3 trillion	$1.016 trillion (roughly 1/3 of the Russell 2000)
Representative Company	Alphabet (3.6% S&P 500 weight)	OpenAI (if listed, roughly 25th largest S&P 500 weight)
Business Model	Asset-light, high margin, zero-cost data	Asset-heavy, paid data, subscription-based
Competitive Arena	Search, social, cloud computing	AI search, video generation, cloud storage

Companies/Assets Involved

• Public Mag 7 (Bearish):
• Alphabet: Core threat from OpenAI's ChatGPT and Sora, potentially eroding ad keyword revenue (core of its profits). Weight 3.6%, larger than the entire energy sector.
• Amazon: OpenAI's cloud storage needs could divert AWS business (which contributes over 100% of profits).
• Meta Platforms: Sora threatens Instagram/Facebook's ad model; already paid $30 million in copyright damages.
• Microsoft: While investing in OpenAI, the Stargate project could weaken its cloud business position.
• NVIDIA: Benefiting from AI hardware demand but its risks are not directly discussed.
• Private Mag 7 (Bullish):
• OpenAI: Valued at $325 billion, ChatGPT and Sora directly threaten the ad models of Google, YouTube, and Meta.
• SpaceX: Valued at $460 billion, competitive path not analyzed in detail.
• Scale AI: Annual subscription fee of $93,000, highlighting the rising trend in data costs.
• Stargate Project: $500 billion investment, potentially reshaping the cloud computing landscape.

Investment Implications

• Short/Underweight Public Mag 7: Especially Alphabet and Meta, as their core ad revenue faces structural erosion. The author implies their high margins are unsustainable, and the asset model shift will compress free cash flow.
• Monitor IPO Opportunities in Private Mag 7: Once companies like OpenAI go public, they could trigger index weight rebalancing and accelerate valuation corrections for the Public Mag 7.
• Beware of "Concentration Risk" in Index Investing: The S&P 500's over-reliance on the Mag 7 could amplify systemic risk, making active management potentially more effective.

Hidden Alpha in Japan's Market: The Deep Value of Entrepreneur CEO Equity Sectors

1. Historical Retrospect: An Overlooked "Outlier" Excess Return

After the internet bubble burst in December 2000, Japan's stock market entered a 12-year negative return cycle, only barely breaking even by mid-2013. Over the same period, the S&P 500 rose about 50%. However, a qualitative sector within Japan's market—entrepreneur CEO (Owner-Operator) companies—achieved a cumulative return of 100%. This reveals two key facts:

• This sector has a very low correlation with Japan's overall market (TOPIX, near zero), offering a rare source of absolute return diversification.
• The excess return did not stem from size or valuation factors but purely from a governance structure where management and shareholder interests are deeply aligned.

2. Current Valuation Mismatch in Japan's Entrepreneur CEO Sector

As of end-2024, the sector's average P/E discount relative to Japan's large-cap market remained at 30-40% (vs. the Nikkei 225's approximately 15x), with many having P/B ratios below 1.0x. This discount does not reflect fundamental deterioration but stems from global institutional investors' habit of "index allocation" to Japan (passive flows into TOPIX/Nikkei 225 components) and cognitive blind spots regarding small to mid-cap entrepreneur companies. Comparative data:

Metric	Japan Entrepreneur CEO Companies	Japan Large-Cap (Nikkei 225)	US Large-Cap (S&P 500)
Average P/E (2024)	9.5x	15.2x	22.5x
Average P/B	0.8x	1.3x	4.2x
ROE	12.8%	9.1%	18.5%
Dividend Yield	2.9%	2.1%	1.4%
Insider Ownership	30-60%	<5%	<3%

Key Contradiction: Entrepreneur CEO companies' ROE is already close to US large-cap levels (12.8% vs 18.5%), yet their valuation is only a quarter of the latter. High insider ownership reduces agency costs, so they should command a premium, not a discount. The roots of this mismatch:

• Liquidity Bias: Most entrepreneur companies have market caps below $5 billion, excluded by large passive funds.
• Analyst Coverage Gap: About 60% of Tokyo Stock Exchange listed companies have no sell-side analyst coverage, especially entrepreneur firms.
• Historical Stigma: Global investors' stereotyping of Japan's "zombie companies" has not fully faded, ignoring governance improvements in good entrepreneur firms.

3. Tokyo Stock Exchange Reforms: Catalysts Are in Place

In 2023, the Tokyo Stock Exchange launched "Market Reforms", requiring listed companies to publicly disclose capital costs, ROE improvement plans, and encouraging cross-shareholding unwinding. This policy strongly favors entrepreneur CEO companies:

• They naturally have high ROE and capital discipline, meeting requirements without forced reforms, and may be repriced due to disclosure.
• The unwinding of cross-shareholdings releases many undervalued assets; entrepreneur companies are more likely to return capital to shareholders via buybacks or special dividends.
• Japan's government-driven NISA reforms attract retail inflows, and retail investors prefer local entrepreneur stories.

4. Comparison with AI Competition Logic: Japan's Entrepreneur Sector is "Anti-Fragile"

The earlier discussion of AI's threat to the Mag 7 (rising data costs, search decentralization, knowledge worker replacement) is essentially disruptive competition. In contrast, Japan's entrepreneur CEO sector benefits from entirely different drivers:

• No Reliance on Data Monopoly: Most operate in manufacturing, precision components, specialty materials, etc.—real economy sectors where AI cannot instantly replace their technical barriers and client relationships.
• Family-Controlled Long-Termism: Founders or family CEOs are more willing to invest counter-cyclically in R&D rather than chase quarterly profits, a relative advantage amid global volatility.
• Yen Depreciation Dividend: Many Japanese entrepreneur companies are export-oriented; yen depreciation directly boosts USD-denominated profits, while Mag 7's revenue is more tied to US domestic consumption.

5. A Reference Case: Lessons from 2000-2013

From December 2000 to June 2013, even with the Bank of Japan at zero rates and deflation persisting, entrepreneur CEO companies achieved excess returns through:

• Heavy share buybacks (insiders knew the stock was undervalued)
• Rejecting diversification, focusing on core niche markets (e.g., robot joint bearings, semiconductor testing equipment)
• Accumulating cash by exploiting Japan's long low-rate environment (cash/market cap ratios often exceeding 50%)

Similar conditions are re-emerging: Japan's interest rate is still among the lowest globally (approx. 0.25%), entrepreneur companies' cash reserves are at record highs, and market sentiment has turned conservative due to Trump tariffs and China's overcapacity. This is precisely a window of opportunity for high-conviction discount assets.

6. Summary: An Undefined "Asset Class"

Japan's entrepreneur CEO sector does not fit traditional industry, value, or growth classifications, and is thus completely overlooked by mainstream asset allocation models. But its existence challenges assumptions of "market efficiency" and "global diversification": within a seemingly low-growth, high-debt sovereign market lies a "niche capital market" of high ROE, low valuation, and low correlation. For investors seeking to avoid Mag 7 concentration risk and find true alpha, it may be the only long-term arbitrage opportunity that does not rely on AI narratives or macro bets.