The author argues that the market is at a "quant-friendly" extreme: U.S. equities and the AI narrative are historically stretched in valuation, with contrarian opportunities concentrated in real assets and low-valuation markets discarded by indices—stance: [cautious/contrarian].
The author likens himself to Copernicus, with the core belief that independent analysis can exploit market inefficiencies to generate excess returns. The original text reads: "We chose our eponym due to his willingness to trust his own analysis when it was dangerously unpopular to do so." In other words: "We chose this name because he was willing to trust his own analysis even when doing so was dangerously unpopular." The author also notes that Copernicus was not only an astronomer but also proposed the quantity theory of money and Gresham's law, implying that the firm's philosophy is rooted in challenging orthodoxy. The article then makes an explicit judgment: mature investors who think independently and trust their own analysis and instincts can generate significant excess returns from market inefficiencies caused by "flawed professional and academic theories and practices." This passage is a founder-perspective narrative with self-mythologizing and marketing overtones; readers should note that this is the viewpoint of a position holder.
As of June 30, 2026, Kopernik managed and advised approximately $9.91 billion in assets. Founded in 2013, it is 100% employee-owned, has 49 employees, and is headquartered in Tampa, Florida. The firm positions itself as a global value equity manager and claims a 40-year track record of philosophy and process. By asset composition, mutual funds account for 41%, UCITS (sub-advisory) 28%, separate accounts 19%, private funds 9%, collective investment trusts 1%, and advisory-only assets 2%. On the leadership team, founder David Iben serves as Co-CIO and Chief Portfolio Manager, and Alissa Corcoran serves as Co-CIO and Director of Research; both hold the CFA designation. This section provides institutional background and contains no specific investment judgments.
The introduction poses the question "Will markets expand forever?" and notes that "definitions of bubbles vary," while also mentioning Howard Marks. At the bottom of the page, three short lines are listed: Will markets expand forever? Definitions of bubbles vary; These questions are easier to answer with Howard Marks. This suggests that the subsequent content may revolve around valuation bubbles, market cycles, and risk; however, this section has not yet elaborated, so no more specific arguments or targets can be extracted.
Howard Marks's checklist is not abstract dogma; its core value lies in providing a workable decision-making framework. The data cited in the continuation injects more brutal empirical support into this framework, but the real analytical focus should be on subtler details.
1. The "Historic Breakout" of the Buffett Indicator and Price to Sales
The continuation explicitly states that both the Buffett Indicator (total market cap/GDP) and the price-to-sales ratio are at "unprecedented" highs. As a thermometer for market sentiment, the price-to-sales ratio (P/S) breaking through its historical extreme of 4.0 is significant. P/S reflects the degree of overextension in long-term earnings power better than P/E, because at cyclical peaks in profit margins, P/E can be inflated, while the linear expansion of P/S directly reflects capital's unconditional trust in "future revenue."
| Indicator | Current Status | Historical Percentile | Implied Meaning |
|---|---|---|---|
| Buffett Indicator | Record high | 100% | Total market cap far exceeds what nominal GDP growth can support |
| S&P 500 P/S Ratio | Approx. 4.0+ | 100% | Revenue multiple overextended, sustained only by ultra-low discount rates |
But what demands even greater vigilance is: high valuation is not only statically "expensive," but dynamically "fragile." When market consensus shifts from "expensive but profitable" to "expensive and still must sell a story," the margin of safety has been completely eliminated.
2. Shiller CAPE and 10-Year Expected Returns: Mathematical Certainty
The continuation cites Hussman Advisors data, noting that based on 75 years of data, current valuations correspond to negative returns over the next 10 years. This is not a new view, but combined with the current CAPE reading (roughly the 45-50 range), it can be estimated:
| Scenario | 10-Year Annualized Nominal Return | Necessary Condition |
|---|---|---|
| Optimistic (valuation maintained) | 0% - 2% | Earnings growth must continue to beat expectations |
| Neutral (valuation drifts toward mean) | -2% - 0% | Earnings need to digest valuation |
| Pessimistic (mean reversion) | -5% to -3% | Both earnings and valuation contract |
This mathematical expectation stands in absurd contradiction to current investor behavior: everyone is chasing a market destined for negative returns, yet no one is considering it.
In the sequel, the SpaceX case is the most striking comparison in the entire piece. The table implies an extreme conclusion: the market is pricing “stories,” while fundamentals have been reduced to a footnote.
| Metric ($M) | SpaceX | Amazon | Gap Multiple |
|---|---|---|---|
| Market Cap (6/16/2026) | 2,660,000 | 2,650,000 | Nearly identical |
| Revenue | 19,301 | 742,776 | Amazon revenue is 38.5x SpaceX |
| Net Income | (8,685) | 90,798 | SpaceX loss, Amazon profitable |
| Free Cash Flow | (27,689) | 10,234 | SpaceX negative, Amazon positive |
| Tangible Book Value | 20,127 | 418,465 | Amazon is 20.8x SpaceX |
Key Analysis:
The top 10 company list for 1980–2026 provided in the follow-up report is an excellent textbook example of "survivorship bias." Several core patterns can be distilled from this table:
1. The replacement cycle for leading companies is roughly 10–15 years. In 1980, IBM, AT&T, and Schlumberger; in 1990, Japanese banks dominated the list; in 2000, Cisco, Intel, and Lucent; in 2010, Apple and Google; and today, NVIDIA, Microsoft, and Amazon. Leaders of each decade rarely retain their leadership into the next.
2. Emerging industries eventually become "traditional industries." In 1990, Japanese banks dominated the list (NTT, Bank of Tokyo-Mitsubishi, etc.). After Japan's asset bubble burst in 1995, these companies saw more than 80% of their market value evaporate. The telecom operators of 2000 (NTT DoCoMo, Deutsche Telekom) were the "core technology assets" of their time, but today almost no one regards them as growth stocks. Will today's "seven AI giants" repeat the same mistake in the 2030s? Most likely yes.
3. "Market leadership" itself is a contrarian indicator of mean reversion. The companies bolded in the table as "market leaders spanning at least two generations" — Exxon, IBM, Microsoft, and Apple — have sustained their lead not because of rapid growth, but because of moats in energy, consumer monopoly, and ecosystems, rather than valuation expansion.
Investment implications: The current highest-weight S&P 500 company (NVIDIA, with a market cap exceeding $5 trillion and a weight above 10%) has the highest single-stock weight in history. Even if the AI thesis fully materializes, such concentration means the influence of a single company on the index has degenerated from "diversification" to "lotterization."
The follow-up article cites ShadowStats' 12.1% inflation rate, a sharp challenge to the official CPI. Although ShadowStats' methodology is controversial, its core argument deserves in-depth analysis:
1. Methodological revisions to the official CPI — Since the 1980s, the BLS has repeatedly adjusted the CPI calculation methodology (from explicit substitution to geometric weighting, and then to adjustments for items such as medical costs and owners' equivalent rent). Each adjustment has objectively lowered the inflation reading. If the 1980 compilation method were used, current U.S. inflation would indeed likely be significantly higher than the official reading of 3.5%.
2. The gap between "actual inflation" and "perceived inflation". When the cost of a New York barbecue rises 31% in a year, the public's inflation perception far exceeds official data. This gap is fertile ground for political friction—voter dissatisfaction stands in opposition to policymakers' claims that "inflation is under control."
3. "Second-wave inflation" pressure from fiscal deficits and money supply. Federal debt as a share of GDP has exceeded 130%, while the monetary base has nearly tripled since 2020. More importantly, interest payments on debt have become the fastest-growing mandatory expenditure in the U.S. federal budget (interest payments exceed the defense budget). To maintain bond-market stability, the Federal Reserve must keep interest rates low—which means real interest rates will remain negative. Negative real interest rates almost always lead to a long-term bull market in physical assets such as gold and platinum.
| Indicator | 1980 | 2000 | 2026 |
|---|---|---|---|
| Federal Debt/GDP | 33% | 55% | 130%+ |
| Monetary Base ($ trillion) | ~0.2 | ~0.6 | ~6.2 |
| Official CPI (annual average) | 13.5% | 3.4% | 3.5% |
| Real Interest Rate (10-yr nominal – CPI) | Positive | Positive | Negative (pronounced) |
The core argument is: Gold's rise from its 2020 low (approximately $1600) to its 2025 high (approximately $3500) is not a "safe-haven" narrative, but rather a balance-sheet restructuring driven by monetary credit deterioration and negative real interest rates. That 40% correction is more likely a healthy cleansing of an "overcrowded inflation trade" than a reversal of the bull market.
Based on the charts in the sequel, the analytical framework for platinum can be drawn with great clarity:
1. The platinum-to-gold ratio sits at an extreme historical low. Kapital's notes show platinum trading at only around 60% of gold's price — the discount itself carries two implications: either gold is overvalued, or platinum is undervalued, or both. Since the 1970s, the platinum-to-gold ratio has typically fluctuated between 0.8 and 1.5; a reading below 0.6 today points to extreme dislocation.
2. Supply-side contraction and demand-side resilience. Platinum is used not only in jewelry and investment but also as a core material in automotive catalysts (for gasoline and hybrid vehicles). Even if BEVs grow to 15% (per the sequel's data), the remaining 85% of internal combustion engine plus hybrid vehicles will still require platinum-palladium catalysts. Meanwhile, mine supply has suffered from a decade of low capital expenditure, with virtually no new mine capacity coming online over the next five years — a textbook tight supply-demand balance.
3. The implied repricing path: If platinum returns to 80% of gold's price (i.e., mean reversion), that implies roughly 50%+ upside from current levels. Combined with the operating leverage of platinum mining companies (fixed costs, high price elasticity), the potential returns of the related equities far exceed the metal itself.
This constitutes a classic "contrarian investment + fundamental resonance" setup — the market is abandoning it, while the fundamentals point in exactly the opposite direction.
The Range Resources (natural gas), Ivanhoe Mines (copper), Paladin Energy (uranium), and emerging market farmland listed in the sequel do not simply represent "inflation beneficiaries"; rather, they are dual-driven by long-term supply-side bottlenecks and demand certainty.
| Asset Class | Supply Constraint | Demand Driver |
|---|---|---|
| Natural gas | Shale capex cuts, LNG export growth | Surging electricity consumption from AI data centers |
| Copper | Declining global ore grades, 8-12 year development cycle for new mines | Grid upgrades, EVs, AI computing infrastructure |
| Uranium | Stockpile depletion, constrained supply from Kazakhstan | Nuclear restart is the only reliable baseload power |
| Agricultural land | Global arable land area essentially unchanged since 1960 | Global population still growing, dietary upgrades |
Particularly noteworthy is that the relationship between population growth and arable land area is widely overlooked. Since 1960, per capita arable land area has fallen by more than 50% (from roughly 4,000 square meters to below 2,500 square meters, and is projected to fall to 1,900 square meters by 2030). Land is an extremely irreplaceable asset, and emerging-market farmland (e.g., Cresud at a 60% discount, Gladstone Land at a 79% discount) is effectively being sold at "distressed prices."
The core insight of this section can be summarized as:
1. The market has reached the extreme of being "quant-friendly": concentration, valuations, cash levels, leverage — all point to historical extremes, while "smart money" chooses to ignore the data that, according to the textbooks, should warrant caution.
2. Narrative is not valuation: the SpaceX vs. Amazon comparison perfectly illustrates that capital prices "the future" far more than "the present" — but over the past thirty years, every round of such extreme pricing has ended in mean reversion.
3. "Nobody refuses a new deal" means opportunity lies in the other direction: when everyone is chasing AI and technology, real assets such as platinum, uranium, and agricultural land are instead trading below their intrinsic value, offering investors willing to act contrarily a significant asymmetric risk-reward opportunity.
4. Cycles are not predicted; they are experienced: the 2026 market bears striking similarities to 1999 and 2007 — "this time is different" arguments are everywhere, but ultimately the operating laws of debt, valuation, and liquidity prevail. Smart investors do not need to guess the turning point, but they must ensure they are standing on the right side when it arrives.
Finally, using the market's actual behavior to capture the essence of this section: when most people are so uniformly optimistic about "stories of the future" (AI, space, risk-free code) and so uniformly dismissive of "lessons of the past," what contrarian investors truly should do is start feeling excited about platinum, agricultural land, and commodity companies with genuine earnings power.
The report's title raises the question of whether money is beginning to flow into scarce essential commodities, but the data table itself provides a more precise inference: almost all real assets have lagged money expansion over the past half century or more. This means that no single commodity is overvalued; rather, a huge "price discount" has accumulated between the entire real-asset class and the fiat currency system.
| Asset | 1960–2026 Increase | Multiple Gap vs. USD Money Expansion | 1973–2026 Increase | Multiple Gap vs. USD Money Expansion |
|---|---|---|---|---|
| Money supply | 134x | — | 71x | — |
| Gold | 114x | 0.85x | 41x | 0.58x |
| Silver | 65x | 0.49x | 23x | 0.32x |
| Farmland | 36x | 0.27x | 16x | 0.23x |
| Oil | 23x | 0.17x | 18x | 0.25x |
| Copper | 20x | 0.15x | 10x | 0.14x |
| Platinum | 19x | 0.14x | 10x | 0.14x |
| Potash | 14x | 0.10x | 9x | 0.13x |
| Timber | 0x* | — | 7x | 0.10x |
*Note: No data for 1960; 1973 marks the complete collapse of the Bretton Woods system and the global transition to a pure fiat currency era.
Key observation: potash is one of the lowest "elasticity coefficients" in the table (0.10–0.13x). This means that since 1960, the absolute price increase of potash has been only one-tenth the magnitude of money expansion. Compared with gold (0.85x), another essential commodity, potash's inflation-hedging characteristics have been systematically ignored. On the demand side, however, elasticity is entirely absent — population growth, upgrading food consumption, and soil potassium depletion form a combination of "rigid demand, extremely low monetary elasticity."
Going deeper: in the half century after 1973, money supply expanded 71x, while farmland (16x), copper (10x), and potash (9x) all rose by less than one-quarter of that. Kopernik's "Cantillon candidates" are essentially a mean-reversion bet: either the value of the fiat currency system is re-anchored, or real commodity prices converge toward the increase in money supply. Potash's brief surge to $1,200+/MT in 2021/2022 and subsequent pullback precisely shows that the market has not yet completed this repricing process.
The "MSCI EM Weight Distribution" table in the report reveals a structural distortion in the passive-investing framework — one that is even more extreme in EM than in developed markets, because EM index construction is weighted by free-float market capitalization, which is concentrated in a few export-oriented technology economies, badly misaligned with economic size.
| Economy/Region | Population | GDP (US$ trillion) | Land Area (million km²) | MSCI EM Weight | YTD Performance |
|---|---|---|---|---|---|
| Korea + Taiwan | 0.08bn | 7.9 (approx. 2.9+5.0) | 0.3 (approx. 0.1+0.2) | 48% | +88%/+56% |
| China | 1.41bn | 18.7 | 9.6 | 21% | -11% |
| India | 1.48bn | 4.2 | 8.4 | 11% | -12% |
| Brazil | 0.23bn | 2.6 | 2.9 | 4% | +14% |
Korea and Taiwan, representing only 2.5% of EM's total population, hold close to half the weight; while China, India, and Brazil combined, with 3.1bn people (about 40% of the global population), have a combined weight of just 36%. The MSCI EM index has effectively been transformed into an "Asia tech export index": three economies (Korea, Taiwan, China) absorb about 69% of the weight, and the three largest individual stocks account for 31% — concentration already higher than that of the top three S&P 500 constituents (around 20%).
This structure leads to two direct consequences:
First, the misallocation of passive capital flows. Global passive money flooding into MSCI EM is effectively providing liquidity to the tech stocks of MSCI Taiwan (32.7x P/E) and MSCI Korea (22.5x P/E), while low-valuation markets such as China (13.0x P/E), Brazil (9.5x P/E), and Indonesia (9.5x P/E) attract limited inflows because of their small weights. The more the index rises, the more concentrated the weights; the more concentrated the weights, the higher the valuations — a self-reinforcing loop.
Second, the so-called "EM beta" has disappeared. Buying the EM index no longer means buying emerging-market growth; it means buying the growth of three tech giants. For active investors, this means the benchmark "anchor" has failed, and the markets excluded or underweighted by the index are precisely the areas with the densest mispricing. This explains why Kopernik, in its Q2 trading, bought heavily into "non-weight" country names such as the Philippines (Ayala, GT Capital), Indonesia (Indofood, United Tractors), and Kazakhstan (Halyk Savings).
The report juxtaposes valuation comparisons between Chinese companies and their U.S. peers. Its essence lies not in simply comparing numbers, but in the fact that each pair of companies in the comparison group operates in almost exactly the same sub-industry — an income statement comparison with variables outside "China vs. U.S." minimized:
| Comparison Group | Chinese Company | P/B | P/E | Dividend Yield | BVPS+DPS Growth | U.S. Peer | P/B | P/E | Dividend Yield | BVPS+DPS Growth |
|---|---|---|---|---|---|---|---|---|---|---|
| Pharmaceutical Distribution | Sinopharm Group | 0.6x | 7.1x | 4.1% | 185% | McKesson | N/A* | 21.9x | 0.4% | Negative |
| Nuclear Power | CGN Power | 1.1x | 13.2x | 3.4% | 188% | Constellation Energy | 7.6x | 21.9x | 0.4% | 147% |
| Telecom Infrastructure | China Communications Services | 0.6x | 7.2x | 5.7% | 123% | MasTec | 5.3x | 57.6x | 0.0% | 215% |
*McKesson's book value is negative due to substantial goodwill and intangible assets, so P/B cannot be calculated.
Note two data points:
First, McKesson's BVPS+DPS is negative. That is, the largest U.S. pharmaceutical distributor has consumed book value through buybacks over the past decade, with cumulative shareholder returns relying mainly on share price appreciation rather than the accumulation of intrinsic value — while Sinopharm Group has achieved 185% cumulative growth in book value per share plus dividends, yet the market gives it only a 0.6x P/B. This points to a deeper fact: U.S. stock price growth has essentially detached from asset creation, while the asset creation of Chinese companies has not yet been acknowledged in prices.
Second, MasTec's P/E is as high as 57.6x. As a telecom infrastructure contractor, its business is highly homogeneous with China Communications Services (telecom infrastructure deployment and maintenance). China Communications Services, with a 5.7% dividend yield and 123% BVPS+DPS growth, trades at only 7.2x P/E; MasTec's growth is indeed faster (215%), but the market pays an 8x P/E premium for that additional growth. It is less about paying for growth than paying for the "U.S." geographic label.
Kopernik explicitly states in the report that it applies a conservative 50% discount to theoretical valuations — which means that even comparing Sinopharm Group's 7.1x P/E with McKesson's 21.9x, Chinese companies already imply a "country risk discount" of more than 60%, and the market discounts further on top of that. What active investors demand is not the elimination of risk, but that risk is already reflected in the price.
Re-reading Q2 trading through the preceding logic reveals three clear execution directions:
| Stock | Quarterly Activity | Market | P/B | P/E | Dividend Yield |
|---|---|---|---|---|---|
| Ayala Corp | Added in April | Philippines | 0.6x | 5.4x | 2.0% |
| GT Capital Holdings | Added in May and June | Philippines | — | — | — |
| First Pacific Co | Added in April and May | Indonesia/Hong Kong | — | — | — |
| LG Corp | Added in April / Trimmed in May and June | Korea | 0.4x | 30.7x | 3.8% |
| Empresas Copec | Added in April and June | Chile | 0.7x | 9.3x | 2.5% |
| PT United Tractors | Added in May | Indonesia | 1.1x | 7.8x | 5.6% |
The common feature of these names: P/B at or below 1x, and systematically ignored by index capital in their respective markets. LG Corp's P/E looks high at 30.7x, but as a holding company, its earnings denominator is depressed by minority interests in subsidiaries; the correct assessment should be based on P/B — 0.4x means the market prices its net asset value at a 60% discount. Holding-company discount + emerging-market discount + passive capital abstention constitute a triple stacking of mispricing.
June saw concentrated additions in precious metals (Royal Gold, Valor Gold, Valterra Platinum), fertilizers (Mosaic, Nutrien), platinum-group metals (Impala Platinum, Sibanye Stillwater), and North American natural gas (Birchcliff, Expand Energy, Range Resources). These are precisely the categories ranked lowest in the "monetary elasticity coefficient" table, with the most room for catch-up — when money supply has expanded 134x while potash has risen only 14x, Kopernik's choice is to buy potash producers (Mosaic, Nutrien) rather than potash futures, using corporate earnings to capture the price repricing.
In April, the report bought S&P 500 Index Puts while continuing to add to China/EM assets. These are two sides of the same coin: in the first half (U.S.), options hedge valuation-shrinkage risk; in the second half (EM/real assets), low-valuation individual stocks provide downside protection. In contrast, June sales of Centene, Humana, and Molina (U.S. health insurance/healthcare services) and of Franklin Resources and Amundi (asset managers) suggest that the "bear market trigger" Kopernik anticipates is not a real-economy recession, but valuation shrinkage within financial assets as interest rates remain at higher levels.
The core inference chain of this entire Introduction is: money supply expanded 134x while real commodities lagged far behind → passive capital monopolized pricing power and structurally distorted the EM index → Chinese assets already embed more than a 50% risk discount yet remain rejected → therefore, the best opportunities for active investors lie in companies that are excluded from indices and labeled "risky" by the market, but whose underlying assets are still creating value. The Q2 transaction details are not a random collection but the gradual position-building of this logic.
The data density in this section is noticeably higher; nearly every line reiterates the same point in numbers: Kopernik is not "optimizing the benchmark" but "avoiding the benchmark." The quantitative evidence from the three model portfolio characteristics deserves closer reading than the position-change list itself.
The valuation metrics for the three strategies uniformly adopt the "aggregate valuation approach" (i.e., total portfolio security value relative to aggregate GAAP/IFRS financial metrics, including companies with negative metrics). Under this approach, the most striking figure is the free cash flow multiple:
| Strategy | P/FCF | Implied FCF Yield (1/P/FCF) | Portfolio P/B |
|---|---|---|---|
| Kopernik International | 2.22 | ≈45% | 0.89 |
| Kopernik Global All-Cap | 3.57 | ≈28% | 0.87 |
| Kopernik Global Opportunities | 5.76 | ≈17% | ≈1.0 |
For reference, MSCI ACWI ex US has a P/FCF of 19.60x. Kopernik International's P/FCF is about 89% lower than the index, and its implied free cash flow yield is nearly nine times that of the index. Even more notable is P/B: both the International and Global All-Cap portfolios have P/B below 1 (0.89 and 0.87, respectively), meaning the portfolios are trading below book value overall; P/TBV is also only about 1.1x, paying almost no premium to tangible asset book value.
Of course, this calculation has a technical feature that cannot be ignored — it includes negative-free-cash-flow companies in the denominator, and resource stocks are precisely prone to accounting losses or negative cash flow at cyclical troughs, which artificially depresses P/FCF. Therefore, 2.22x is best interpreted as evidence of how extreme the market's pricing assumptions for these assets have become, rather than a precise promise of future returns. But even after removing that distortion, P/B below 0.9x and P/TBV of about 1.1x still constitute a margin of safety independent of the earnings cycle.
Market-cap distribution is another extremely strong signal of deviation.
| Metric (US$ million) | Kopernik International | MSCI ACWI ex US |
|---|---|---|
| Weighted average market cap | 15,098 | 276,240 |
| Median market cap | 6,331 | 39,584 |
| Large-cap (>$10B) weight | 29.65% | 94.58% |
| Mid-cap ($2B–$10B) weight | 37.75% | 5.40% |
| Small-cap (<$2B) weight | 6.10% | 0.02% |
| Metric (US$ million) | Kopernik Global All-Cap | MSCI ACWI |
|---|---|---|
| Weighted average market cap | 19,141 | 986,948 |
| Median market cap | 3,148 | 28,436 |
| Large-cap (>$10B) weight | 27.00% | 98.01% |
| Mid-cap ($2B–$10B) weight | 38.85% | 1.99% |
| Small-cap (<$2B) weight | 23.95% | 0.01% |
Kopernik International's weighted average market cap is approximately 5.5% of the index; Global All-Cap is even more extreme, at about 1.9% of the index. In terms of median market cap, Global All-Cap's median is only $3.1 billion, roughly nine times smaller than the index's $28.4 billion.
Global All-Cap's market-cap structure is especially worth examining: it holds both a 27% large-cap position and nearly 24% in small caps. The weighted average market cap ($19.1 billion) is lifted by a few large positions, while the median market cap ($3.1 billion) reveals its true home — small- and mid-cap companies with low attention and low analyst coverage. This structure of "concentrated head, extremely wide tail" means the portfolio's returns are necessarily highly dependent on whether a few forgotten assets can be repriced.
| Metric | Kopernik International | Kopernik Global All-Cap |
|---|---|---|
| Active Share | 97.87% | 99.13% |
| Non-index holdings weight | 53.95% (41/76) | 68.60% (83/121) |
Active Share above 97% means the portfolio's overlap with the benchmark is minimal; Global All-Cap's 99.13% is even closer to a "hand-picked list" entirely independent of MSCI ACWI. The proportion of non-index holdings is equally striking: of the International portfolio's 76 holdings, 41 are not in MSCI ACWI ex US at all; of the Global All-Cap portfolio's 121 holdings, 83 are not in MSCI ACWI.
These figures show that Kopernik is not making overweight/underweight judgments within an index framework, but selecting stocks independently outside the index's coverage. Non-index holdings abound in small caps, emerging markets, and frontier markets, which precisely explains why the portfolio's market-cap distribution is so disconnected from the index.
| Region | International | MSCI ACWI ex US | Global All-Cap | MSCI ACWI |
|---|---|---|---|---|
| United States | 2.10% | 0.00% | 10.15% | 63.63% |
| Non-U.S. | 71.40% | 100.00% | 79.65% | 36.37% |
| Developed markets | 39.30% | 66.37% | 50.95% | 87.77% |
| Emerging markets | 34.20% | 33.63% | 38.85% | 12.23% |
MSCI ACWI's U.S. weight is as high as 63.63%, while Global All-Cap allocates only 10.15%, simultaneously overweighting emerging markets by about 26.6 percentage points. Although the International portfolio is benchmarked to ACWI ex US and theoretically should not hold U.S. stocks, it still contains a 2.10% U.S. position — indicating that Kopernik does not mechanically exclude countries based on the benchmark; when U.S. assets are cheap enough, they can also be included.
Another detail: the sum of "U.S. + Non-U.S." in the two data sets does not reach 100%. Taking International as an example, 71.40% + 2.10% = 73.50%. The remaining difference mainly corresponds to two parts: first, Russian securities explicitly excluded from the characteristic calculations; and second, about 1% in option positions and cash. This means the portfolio's actual risk exposure is more complex than the table presents.
The top ten holdings of International and Global All-Cap show a high degree of consistency, with the core holding circles heavily overlapping:
| Holding Theme | Representative Companies |
|---|---|
| Platinum-group metals | Valterra Platinum, Impala Platinum |
| Gold/Copper | Seabridge Gold, Novagold Resources, Ivanhoe Mines |
| Potash/Basic materials | K+S, Glencore |
| Telecom | LG Uplus, KT Corp |
| Agriculture | Golden Agri-Resources |
The top ten holdings account for approximately 27.00% of the International portfolio and approximately 25.50% of Global All-Cap. This may look diversified, but given the exceptionally high Active Share, this roughly one-quarter of the book actually carries the vast majority of the portfolio's risk exposure.
Valterra Platinum Ltd is the most noteworthy "new face" in this data — it appears simultaneously as the largest holding in the International portfolio (4.50%) and Global All-Cap (4.00%). As a South African platinum company, it joins Impala Platinum to form roughly a 7% weight in the platinum sector (measured in the International portfolio). Placing such a high weight on platinum-group metals is a bet on a niche segment that mainstream capital has long avoided but where supply-side constraints are clear.
Another observable feature: telecom holdings (LG Uplus, KT Corp) appear in both portfolios at the same time, with combined weights of approximately 4.75% each. Such assets are typically seen as growth-challenged in developed markets, but in the context of emerging markets/non-U.S. developed markets, they offer stable free cash flow and buyback capacity, closely matching Kopernik's "low valuation + cash flow" preference.
An easily overlooked detail is that the Global Opportunities portfolio's characteristic data has been disclosed, but the original text explicitly states that the strategy "has not yet been launched or fully implemented" — meaning it has not been launched or fully deployed.
Its outline is already clear, however: P/FCF is 5.76x, significantly higher than International (2.22) and Global All-Cap (3.57), but still far lower than mainstream indices; P/B is about 1.0x; median market cap is roughly $10 billion; and weighted average market cap is close to Global All-Cap (about $19.2 billion), but the median is considerably larger. This suggests that Kopernik's Global Opportunities strategy may be aimed at clients willing to accept a somewhat higher free cash flow multiple and preferring larger companies — a middle option between "extreme deep value" (International/Global All-Cap) and more conventional value investing.
Finally, two footnotes that readers can easily overlook need to be noted.
First, all characteristic data explicitly excludes Russian securities, but the wording is "held in the representative portfolio" — that is, the representative portfolio still holds Russian securities, but they are excluded only in the characteristic calculations. This indicates that Kopernik's treatment of Russian assets is not a simple "zeroing out," but rather maintaining the holdings while segregating them in information disclosure and portfolio characteristic presentation. For investors, this means the actual portfolio's geopolitical risk is higher than the tables show.
Second, all of these are "Model Portfolio" data, not audited performance records. Global Opportunities has not even gone live yet. Combined with the aggregate valuation method's inclusion of negative-metric companies, the absolute levels of P/FCF, P/E, and other readings may be distorted by loss-making companies at cyclical troughs. Therefore, these data are best treated as a "quantitative snapshot of Kopernik's strategic intent," not a promise of future returns.
The following is the additional analysis for Part 5/7 of the sequel, focusing on dimensions such as the extremity of portfolio construction, historical win-rate structure, the mathematical implications of IRR, and fee governance.
An Active Share of 98.98% means the portfolio overlaps with MSCI ACWI by only about 1%. This is not a "high-active" fund in the traditional 60–80% Active Share sense, but a portfolio constructed almost entirely independently of the index.
| Metric | Typical Active Fund | Kopernik Global Opportunities |
|---|---|---|
| Active Share | 60-80% | 98.98% |
| Non-index position weight | 5-20% | 39.68% |
| Annual tracking error (estimated) | 3-6% | 15-25% |
Notable detail: The 39.68% non-index weight corresponds to 25 of 63 holdings not covered by MSCI ACWI, and this already excludes non-equity securities. This means nearly 40% of the portfolio sits in an "index blind spot"—unable to contribute relative returns when the index surges, but potentially offering an independent source of returns when the index declines or styles rotate. This also explains why the portfolio was able to beat its benchmark by 24 percentage points in 2025: when "hidden assets" such as platinum trusts and small-cap mining stocks were repriced, funds holding them captured pure alpha with no offsetting beta.
The cost, meanwhile, is the sharp drawdown in Q2 2026: QTD -5.93% (international strategy) against a benchmark of +14.49%, a relative underperformance of roughly 20 percentage points in a single quarter. This is the normal electrocardiogram of an extreme active strategy, not an abnormal signal. Investors who cannot tolerate significant underperformance lasting three quarters or more are fundamentally unsuited to holding a strategy with Active Share above 90%.
The GIPS report provides complete year-by-year performance, making it easy to identify the strategy's style cycles:
| Year | Gross (%) | Benchmark (%) | Excess (%) | Market Style |
|---|---|---|---|---|
| 2015* | -10.18 | -9.32 | -0.86 | Global turmoil |
| 2016 | +27.55 | +4.50 | +23.05 | Value/commodities rebound |
| 2017 | +11.61 | +27.19 | -15.58 | Global growth/U.S. tech |
| 2018 | -5.29 | -14.20 | +8.91 | Risk-off/deleveraging |
| 2019 | +17.60 | +21.51 | -3.91 | Growth led |
| 2020 | +20.50 | +10.65 | +9.85 | Post-pandemic V-shape/gold |
| 2021 | +18.07 | +7.82 | +10.25 | Cyclicals/value revival |
| 2022 | -12.85 | -16.00 | +3.15 | Inflation/rate hikes |
| 2023 | +15.68 | +15.62 | +0.06 | AI divergence |
| 2024 | -2.93 | +5.53 | -8.46 | Tech concentration |
| 2025 | +56.58 | +32.39 | +24.19 | Real assets/platinum |
Statistical results:
Regularity identified: The losing years (2017, 2019, 2024) all coincided with one-sided rallies in global growth/tech stocks in those years; the winning years covered down markets, range-bound markets, inflationary markets, and real-asset bull markets. This asymmetric structure—outperforming substantially with the wind and underperforming mildly against it—is precisely the mathematical foundation of deep-value strategies' long-term effectiveness. The current relative drawdown of -20 percentage points in Q2 2026 resembles the 2024 pattern; historical experience points to a possible compensatory rebound within 12-24 months.
The table presents a concise yet powerful logic: if the potential upside from current holding prices is 50%/100%/150%, the corresponding annualized internal rates of return over holding periods of 1–10 years are as follows:
| Holding Period | Upside 50% | Upside 100% | Upside 150% |
|---|---|---|---|
| 1 Year | 50.0% | 100.0% | 150.0% |
| 3 Years | 14.5% | 26.0% | 35.7% |
| 5 Years | 8.5% | 14.9% | 20.1% |
| 10 Years | 4.1% | 7.2% | 9.6% |
Key Insight: The table's implicit message is not a "short-term spike" but an "anchor of long-term compounding." Even with potential upside as high as 150%, if it is realized gradually over 10 years, the annualized return is only 9.6%—far below many investors' intuitive expectation of a "doubling." Conversely, even under the most conservative 50% upside assumption, the 10-year annualized return of 4.1% remains significantly higher than current developed-market government bond yields (assuming a 2026 environment), and the portfolio also retains the inflation-protection attributes of real assets.
Link to 2025 Performance: Even after a +56.58% annual surge, the fund manager still provides a potential upside of 50%–150%, indicating that they view the 2025 rally as the starting point of discount repair rather than its endpoint. If real assets such as platinum, uranium, and potash are in the early stages of long-term supply-demand imbalances, this calculation is logically self-consistent.
A precise breakdown of the top ten positions:
| Asset Class | Holdings | Total Weight |
|---|---|---|
| Platinum group metals | Sprott Phys Plat (7.0%) + Valterra (5.0%) + Impala (4.0%) | 16.0% |
| Gold | Seabridge Gold | 3.0% |
| Fertilizer/Potash | Nutrien | 4.0% |
| Natural gas | Range Resources | 2.75% |
| Copper/Base metals | Ivanhoe Mines | 2.5% |
| Subtotal: top 10 physical assets | 28.25% |
Including the Sprott trust's direct holdings of physical platinum/palladium, roughly 30% of the portfolio's weight is directly tied to physical commodity prices. Platinum group metals alone account for 16%, constituting the portfolio's largest risk exposure.
The logic supporting this concentration: Platinum faces dual drivers of South African supply contraction (power shortages, rising deep-mine costs) and structurally rising demand (hydrogen fuel cells, substitution for palladium in gasoline vehicle catalysts); palladium has a persistent supply deficit. But the risks are equally significant—if a global recession reduces auto production and sales, or if the hydrogen technology roadmap underperforms expectations, platinum prices could remain depressed for an extended period. A 16% single-commodity position could, in extreme cases, contribute a -20% to -30% drag on the portfolio.
The two strategies share the same investment philosophy, but differ in allocation. Their performance divergence in recent years provides an interesting comparison:
| Metric | International (Gross) | MSCI ACWI ex US | Global All-Cap (Gross) | MSCI ACWI |
|---|---|---|---|---|
| 1-Year | +20.47% | +27.66% | +25.69% | +23.67% |
| 3-Year | +18.98% | +18.80% | +24.47% | +19.68% |
| 5-Year | +9.60% | +8.78% | +13.29% | +10.98% |
| 10-Year | +10.53% | +9.92% | +14.55% | +12.78% |
| Since Inception | +10.95% | +7.91% | +11.63% | +11.18% |
Core differences: Global All-Cap outperformed over the 1/3/5/10-year periods, while International only barely matched over the 3-year horizon and significantly underperformed over the 1-year period. There may be three reasons:
Implications for investors: For those who favor non-US markets, especially emerging markets and real assets, International is the purer choice; for those who wish to execute a contrarian strategy on a global basis and reduce short-term relative volatility, Global All-Cap has a more stable track record. The two strategies' return gap of nearly 4 percentage points over the past 10 years is a difference in choice that cannot be ignored.
The extreme divergence in Q2 2026 — a portfolio return of approximately -5.9% versus a benchmark return of approximately +14.5% — needs to be viewed over a longer cycle:
| Historical quarters of major underperformance | Actual performance over the following 12 months | Implication |
|---|---|---|
| 2017 full-year underperformance of 15.6% | Outperformed by 8.9% in 2018 | Underperformance is often followed by a reversal |
| 2019 underperformance of 3.9% | Outperformed by 9.9% in 2020 | A modest gap is followed by substantial compensation |
| 2024 underperformance of 8.5% | Outperformed by 24.2% in 2025 | The strongest underperformance is followed by the strongest rebound |
| Q2 2026 relative underperformance of approximately 20% | To be observed | Historical patterns point to the next 12–24 months |
Of course, history does not simply repeat itself. But if the fundamentals of the portfolio's holdings — the platinum supply-demand gap, undervaluation of Korean telecom stocks, and the natural gas cycle — have not deteriorated, then the short-term relative drawdown actually raises the potential upside. That is a realistic footnote to the 50%–150% potential upside shown in the table.
The fee examples in the material reveal the true impact of management fees in a compounding environment:
| Scenario | 10-Year Terminal Value | Annualized Return |
|---|---|---|
| No fees (assuming 10% annualized) | $270,704 | 10.47% |
| After deducting 0.90% annual fee | $247,581 | 9.49% |
The fee difference of $23,123 represents 8.54% of the gross terminal value. Against the GIPS net return gap, actual total fees (including transaction costs) amount to approximately 95-131 bp per year, a mid-range level within the industry.
Positive governance factors:
Among the top ten holdings, Korean companies account for a combined 9.25% weight (KT Corp 3.5% + LG Uplus 3.25% + LG Corp 2.5%). This is not a random selection but a concrete elaboration of the "quality emerging-market companies" thesis:
The Korean market exhibits a pronounced "Korea Discount" (foreign investors' governance concerns, chaebol structures, geopolitical risks), causing these companies to be valued below their global peers. This is precisely the soil that contrarian investors need: high-quality companies with predictable cash flows, yet the market discounts them for structural reasons. As South Korea's capital market reforms (benchmarked against the governance improvements at the Tokyo Stock Exchange in Japan) advance, these holdings may undergo a valuation re-rating.
The core arguments added in this section can be summarized as follows:
1. 98.98% Active Share is a double-edged sword — sharp short-term underperformance is proof that the strategy is operating correctly, not that it has failed
2. An 11-year 73% win rate and a 1.37:1 profit/loss ratio confirm the asymmetric structure of the excess returns
3. The analytical implication of the Potential Upside table: 50-150% potential upside corresponds to a 5-year 8.5-20.1% IRR, which is superior to bonds even under conservative assumptions
4. The 16% concentration in platinum group metals shows that the manager is willing to place heavy bets on a single thesis; investors need to assess their own risk tolerance
5. Global All-Cap has delivered better actual results than International; if forced to choose only one, the former offers better long-term risk-adjusted returns
6. Significant underperformance in 2026Q2 has historically often foreshadowed excess compensation over the subsequent 12-24 months, but this is a statistical regularity rather than a guarantee
Following the earlier analysis of the strategy framework, benchmark, and fee structure, the historical performance data in the GIPS GIPS® Report provides a more empirically valuable analytical dimension. This section focuses on the complete disclosure data of the Global All-Cap Composite over more than a decade, revealing its return distribution, risk-adjusted characteristics, and asset evolution logic.
Across the 12 full calendar years from 2014 to 2025, the strategy (gross returns) outperformed MSCI ACWI (Net) in only 5 years (2016, 2020, 2021, 2022, 2025) and underperformed in as many as 7 years (2014, 2015, 2017, 2018, 2019, 2023, 2024). If one looks only at the "win rate," the strategy stands at only about 42%, but this does not reflect the true full picture.
The key difference lies in the highly asymmetric distribution of excess returns:
| Period | Average Excess in Outperformance Years | Average Excess in Underperformance Years |
|---|---|---|
| 2014-2025 | +18.28 percentage points | -8.48 percentage points |
| Excluding 2025 | +9.03 percentage points | -8.48 percentage points |
Even after excluding the record 44.25-percentage-point excess return of 2025, the average advantage in outperformance years still exceeds the average disadvantage in underperformance years. This payoff structure of "winning less often but winning more; losing more often but losing less" is the fundamental source of its long-term cumulative excess returns.
Based on chain-linked compounding of annual returns (NAV of 100 at the beginning of 2014), as of the end of 2025:
This indicates that the single year of 2025 contributed the majority of the strategy's long-term excess returns. Even so, in 2022 the strategy fell only 8.64% while the benchmark fell 18.36%, providing a positive deviation of 9.72 percentage points; this risk asymmetry is equally worthy of attention.
However, the strategy's volatility is significantly higher than the benchmark's. Using a rough calculation based on the 3-year rolling annualized standard deviation (for the years with available data, 2018-2025):
| Metric | Strategy | Benchmark |
|---|---|---|
| Average 3-year annualized standard deviation | 17.77% | 14.77% |
| Highest 3-year annualized standard deviation | 22.86% (2022) | 19.86% (2022) |
| Lowest 3-year annualized standard deviation | 11.05% (2019) | 10.48% (2018) |
The strategy's volatility is on average about 3 percentage points higher than the benchmark's, and in certain years (such as 2020) internal dispersion reached as high as 6.59%, implying notable return differences across accounts. If one roughly measures excess return per unit of risk as annualized excess divided by average standard deviation, the strategy comes to about 0.65 (1.94/3), which is not particularly outstanding. However, given the downside protection observed in bear scenarios (2022, 2023), its downside participation rate is clearly lower than its upside participation rate.
| Year | Strategy Gross Return | Benchmark Return | Net Return | Notes |
|---|---|---|---|---|
| 2022 | -8.64% | -18.36% | -9.29% | Bear-market resilience evident |
| 2020 | +35.99% | +16.25% | +35.15% | Strong rebound, but internal dispersion reached 6.59% |
| 2014 | -18.01% | +4.16% | -18.67% | Largest relative drawdown, 22 percentage points behind benchmark |
| 2015 | -11.74% | -2.36% | -12.37% | Second consecutive year of deep underperformance |
| 2025 | +66.59% | +22.34% | +65.44% | Extreme outperformance year, accounting for approximately 40% of the 12-year cumulative excess returns |
The two consecutive years of significant underperformance in 2014-2015 caused composite assets to shrink at one point to US$800 million (in 2014), but the 53.12% rebound in 2016 quickly recovered the losses and drove assets to approximately US$2.4 billion. This "deep pit - steep climb" trajectory demands exceptional patience from investors and is also the reason the strategy presents high volatility through the 3-year standard deviation in the GIPS report.
This report covers the Global All-Cap strategy, whose benchmark is MSCI ACWI (including the U.S. market). The International strategy mentioned earlier, by contrast, uses MSCI ACWI ex USA as its benchmark, and the international strategy can invest at most 15% of assets in the U.S. The two may exhibit significantly different performance in 2025 — Global All-Cap was able to participate fully in the U.S. market (especially if energy and resource-related stocks performed strongly), while the International strategy was constrained by its U.S. position cap. This explains why the firm needs to issue separate GIPS reports and draw a clear distinction in investment scope.
The report claims to have passed independent verification (covering July 1, 2013 through December 31, 2025) and that the strategy has undergone a dedicated performance review. However, note that:
These details remind users that although the GIPS framework enhances transparency, the strategy's excess returns are largely driven by a very small number of years (2016, 2020, 2022, 2025), and historical performance cannot serve as a guarantee of future returns.
Governance Implications of the Outsourcing Model
Kopernik outsources its middle- and back-office functions (including performance calculation) to SEI Investments Company. This arrangement is not merely a cost optimization; it also carries an implicit independence consideration — separating performance calculation from investment decisions reduces the risk of human intervention. For GIPS compliance, third-party calculation can enhance the credibility of the results. However, note that outsourcing does not transfer ultimate compliance responsibility: Kopernik remains fully responsible for the content of the calculation results and the accuracy of disclosures. If SEI's scope of services includes valuation, then the choice of models for daily security pricing (such as liquidity discounts and the timing of foreign exchange closing rates) will directly affect intraday NAV, while delays in trade matching within settlement systems may cause month-end reconciliation differences — these details are hidden dimensions for evaluating the quality of the outsourcing.
The "Hidden Frictions" in Valuation Basis and Reinvestment Assumptions
Performance is denominated in U.S. dollars and assumes dividend reinvestment, but the following is not specified:
Investors are advised to request the complete valuation policy document from Kopernik (the text states it is available upon request), focusing in particular on the treatment of the "valuation discount" arising from foreign ownership restrictions.
| Fee Model | AUM Tier | Fee Rate | Performance Fee Included | Characteristics |
|---|---|---|---|---|
| Model A (with performance fee) | All sizes | 0.25% | Yes, with lock-up | Low base fee + high elasticity, suited to clients seeking excess returns |
| Model B (management fee only) | 0-$50M | 0.80% | No | High fee rate for small clients, but no performance fee |
| $50-150M | 0.75% | No | Marginal decline | |
| $150-250M | 0.70% | No | ||
| $250-350M | 0.65% | No | ||
| >$350M | 0.60% | No | Large clients enjoy scale discounts |
Key Observations:
1. Cross-subsidization design: clients selecting Model A pay a 0.25% management fee, far below the 0.80% small-tier rate of Model B; however, if the performance fee is triggered, the total fee rate may exceed 1.5% (assuming a 20% performance fee share and excess returns >6%). This in effect allows low-fee clients to provide the scale base for high-fee clients, while partially transferring performance risk to the investment adviser.
2. The meaning of the lock-up: the performance fee lock-up is typically designed to prevent clients from redeeming at performance peaks and subscribing at troughs, which would create unfair accruals; however, the lock-up also reduces client liquidity. For institutional investors, the match between the lock-up period and their own liability duration needs to be assessed.
3. Voluntary fee reduction clause: the text mentions that one pooled account voluntarily lowered its management fee to control the expense ratio. This may cause that account's net fee rate to be lower than that of other comparable accounts, leading the composite's internal dispersion to be understated — because dispersion is calculated based on gross returns, yet fee adjustments may affect the portfolio's cash retention and reinvestment efficiency.
4. Actual fee variation: the text emphasizes that "actual fees may vary depending on client circumstances," meaning scale discounts are not the only variable; historical working relationships, the complexity of investment restrictions, and co-investment terms may all affect the final fee rate. The fee details in Form ADV Part II are worth checking line by line.
Since 2019, accounts with cumulative net cash flow in a single month exceeding 20% of the prior month-end market value have been temporarily removed from the composite. The design of this threshold merits scrutiny:
The risk provisions cover four categories: liquidity, small capitalization, non-U.S. markets (emerging/frontier), and the natural resources sector. Compared with a standard template, two points warrant deeper consideration:
The appendix defines 14 indices/ETFs, including the VanEck Gold Miners ETF (GDX) and the NYSE Arca Gold BUGS Index, reflecting that Kopernik's allocation weight to gold stocks is significantly higher than that of a typical global equity strategy. GDMNTR, which GDX tracks, is a price return index, while BUGS is an equal-weighted index; the differences between the two mean that:
In addition, the constituent counts in the index definitions may suffer from potential update lags:
| Index | Countries/Regions Covered | Number of Constituents | Coverage | Notes |
|---|---|---|---|---|
| MSCI ACWI | 47 | 2,517 | ~85% | Includes A-shares at 20% inclusion ratio |
| MSCI EM | 24 | 1,178 | ~85% | No China A-shares |
| MSCI Europe | 15 | 397 | ~85% | |
| MSCI Brazil | 1 | 46 | ~85% | |
| MSCI Philippines | 1 | 10 | ~85% | Few constituents, high volatility |
| MSCI Indonesia | 1 | 11 | ~85% | |
| MSCI China | 1 | 576 | ~85% | A-shares at 20% |
Here, both MSCI Korea and MSCI Taiwan show 77 constituents, but KOSPI and MSCI Korea differ in definitional scope — KOSPI covers all common stocks, whereas MSCI Korea covers only 85% of large- and mid-cap market value. The return difference between the two can be as high as 5% or more, particularly during periods of regulatory policy changes in the South Korean market. The appendix does not provide the conversion relationship between KOSPI and MSCI Korea; when evaluating the performance of the portfolio's South Korean exposure, investors need to clarify which benchmark is actually being used.
This section appears to be standard compliance disclosure, but it actually reveals the following strategy characteristics:
1. Through the "performance fee + lock-up" structure in its fee arrangements, it attracts long-term capital that has confidence in alpha;
2. Through the large cash flow exclusion mechanism, it maintains the representativeness of reported performance;
3. Through the gold stock index and specific emerging market indices in its benchmark definitions, it signals an allocation style of "contrarian investing + commodity cycle."
For investors, the supplementary documents most worth obtaining are:
The completeness of this disclosure ranks in the upper-middle tier among comparable small- and mid-sized investment advisers, but the "estimated actual management fee" language still needs to be reconciled line by line against the actual fees deducted in the period-end account statements.
| Instrument | Direction | Author's Stance in One Sentence | Key Data |
|---|---|---|---|
| Impala Platinum, Sibanye Stillwater, Valterra Platinum | Add | Platinum's discount is at an extreme historical level; supply-demand is in tight balance, leaving ample room for mean reversion | Platinum/gold ratio approximately 0.6, historical range 0.8–1.5 |
| Royal Gold, Valor Gold | Add | Negative real interest rates and deteriorating monetary credibility drive a re-rating of gold assets | Gold rose from a 2020 low of approximately $1,600 to a 2025 high of approximately $3,500 |
| Mosaic, Nutrien | Add | Potash's money elasticity coefficient is only 0.10–0.13x, demand is rigid, and prices seriously lag money creation | Potash rose 14x over 1960–2026, money supply rose 134x |
| Birchcliff, Expand Energy, Range Resources | Add | North American natural gas supply is contracting while AI data center electricity demand is surging | Shale capex cuts, LNG export growth |
| Ayala Corp | Add | Philippine holding company systematically ignored by index capital, with a triple discount stacked | P/B 0.6x, P/E 5.4x, dividend yield 2.0% |
| GT Capital Holdings | Add | Similar to Ayala, a low-attention name in the Philippine market | Added positions in April, May, and June in succession |
| First Pacific Co | Add | Indonesia/Hong Kong holding company, trading at a discount to net asset value | Added positions in April and May |
| Empresas Copec | Add | Chilean holding company with P/B below 1x | P/B 0.7x, P/E 9.3x, dividend yield 2.5% |
| PT United Tractors | Add | Indonesian heavy equipment distributor, high dividend + low valuation | P/B 1.1x, P/E 7.8x, dividend yield 5.6% |
| LG Corp | Add, then partial reduce | Holding company discount, P/B at only 0.4x, but within the quarter the position was first added then cut | Added in April, reduced in May/June |
| S&P 500 Index Put | New | Using options to hedge tail risk from U.S. equity valuation compression | Bought in April |
| Centene, Humana, Molina | Reduce | Wary of valuation compression within financial assets under high interest rates, rather than an economic recession | Sold in June |
| Franklin Resources, Amundi | Reduce | Asset managers are sensitive to market beta; valuation compression risk is anticipated | Sold in June |
| Sinopharm Group | Hold / Watch | China's pharmaceutical distribution leader; asset creation has not been recognized by the price | P/B 0.6x, P/E 7.1x, dividend yield 4.1%, BVPS+DPS growth 185% |
| CGN Power | Hold / Watch | Scarce nuclear power asset; China's valuation is only a fraction of U.S. peers | P/B 1.1x, P/E 13.2x, dividend yield 3.4%, BVPS+DPS growth 188% |
| China Communications Services | Hold / Watch | Telecom infrastructure business is homogeneous with U.S. MasTec, with a valuation gap of 8x in P/E | P/B 0.6x, P/E 7.2x, dividend yield 5.7% |
| Cresud, Gladstone Land | Hold / Watch | Emerging-market farmland is being sold at "distressed prices" | Cresud at a 60% discount, Gladstone at a 79% discount |
| Ivanhoe Mines | Not disclosed | Copper supply bottleneck plus grid/EV/AI computing demand, with a clear long-term logic | Global ore grades declining; new mine development cycle 8–12 years |
| Paladin Energy | Not disclosed | Nuclear restart drives uranium demand; inventories depleted and supply constrained | Kazakhstan supply constrained |
| SpaceX | Not disclosed | Used as a counterexample: the market is pricing a long-dated option on "Mars colonization" | Market cap $2.66T ≈ Amazon, revenue is 1/38.5 that of the latter, FCF -$27.7B |
| Amazon | Not disclosed | Comparison case: positive earnings and positive FCF, valuation far below SpaceX | Revenue $742.8B, net income $90.8B |
| NVIDIA | Not disclosed | Extreme index concentration, devolving from "diversification" to a "lottery" | Market cap over $5 trillion, S&P 500 weight over 10% |
| McKesson | Not disclosed | U.S. peer comparison: buybacks consume book value; shareholder returns rely on share price rather than intrinsic value | BVPS+DPS growth negative, P/E 21.9x |
| Constellation Energy | Not disclosed | U.S. nuclear peer, valuation far higher than Chinese counterparts | P/B 7.6x, P/E 21.9x |
| MasTec | Not disclosed | U.S. telecom infrastructure peer; the market pays an 8x premium for the "U.S." label | P/E 57.6x, BVPS+DPS growth 215% |
| Cisco | Not disclosed | Historical bubble case: the same compelling narrative ultimately erased over 80% of market cap | 2000 forward P/E over 100x, market cap over $1 trillion |