Hosking Partners is a London boutique founded in 2013 by Jeremy Hosking, a portfolio manager at Marathon Asset Management for over 25 years. It runs a single global equity strategy built on the capital-cycle, supply-side approach — contrarian, long-term, and unusually diversified (350+ holdings) under a multi-counsellor model, managing around $5.5bn.
The market is experiencing one of the biggest momentum rallies in history, with 85% of gains coming from AI-related stocks. Four stocks alone (Nvidia, Micron, Broadcom, Sandisk) are expected to contribute 40% of S&P 500 earnings growth. The author argues this extreme concentration won't last; markets will refocus on 'capital cycle' analysis—looking at which industries face shortages and need new investment, such as memory chips, power, and defense. A contrarian strategy that diversifies across sectors briefly lagged during the Q2 sentiment surge but still beat the global index by 13.5% over the past year. Worth reading for a reminder that supply bottlenecks, not just AI, could be the next big opportunity.
The author believes the market is returning to traditional capital-cycle logic, with short-term exuberance being part of investors relearning fundamentals; however, the author remains cautious about AI-driven index concentration and future equity supply, while expressing confidence in the firm's own contrarian positioning [Cautious].
The strategy generated a net return of 11.7% in Q2 and 15.3% YTD; the ACWI benchmark rebounded 14.9% in Q2 amid "animal spirits," leaving the strategy 3.2 percentage points behind. YTD, however, it still leads by 4.0 points, and over the past 12 months it has outperformed by 13.5 points.
Performance Comparison:
| Period | Strategy (net) | MSCI ACWI | Relative Performance |
|---|---|---|---|
| Q2 2026 | +11.7% | +14.9% | -3.2% |
| YTD | +15.3% | Not disclosed | +4.0% |
| Past 12 months | Not disclosed | Not disclosed | +13.5% |
The author attributes this to the fact that this year has been one of the largest momentum markets in modern stock-market history: AI spending beneficiaries contributed 85% of this year's market return. According to analyst estimates, four stocks—Nvidia, Micron, Broadcom, and Sandisk—will contribute 40 percentage points of S&P 500 earnings growth (source: Empirical Research Partners, "The Anti-AI Plays," June 2026). The author believes that when such a narrow leadership structure dominates an index that is "supposed to be diversified," a genuinely differentiated strategy will struggle to outperform in the short term. Note that this is self-justification from the position-holder's perspective—the author uses "genuine differentiation" to explain the short-term lag, then points to the 12-month track record to demonstrate the strategy's value. In the first half of the year, volatility was extreme under opposing narratives including AI dominance, trillion-dollar IPOs, and geopolitical turmoil; the strategy had outperformed the benchmark by a wide margin in Q1.
The author judges that the shortages and bottlenecks in memory semiconductors, critical minerals, power, and defense are essentially surging demand meeting inelastic short-term supply. Investors are relearning fundamentals, and the market is returning to traditional capital-cycle investing.
Over the past decade, investors predominantly allocated to companies with "light capital, heavy intangible assets" and highly flexible input structures. Now, supply lead times and industry-structure analysis are again critical for judging whether improvements in returns on capital can persist or whether profits will quickly mean-revert. Stance: [cautious]—the author acknowledges that today's frenzy "may be a once-in-a-lifetime phenomenon" but regards it as a return to the capital cycle. The author quotes colleague Michael Godfrey in the Hosking Post, "'Normal' Service Resumes": "The exuberance we see in markets today is an important part of investors re-learning fundamentals." In other words: "The exuberance we see in markets today is an important part of investors re-learning fundamentals." The author adds that capital now has a cost; the lesson is both unavoidable and painful. Capacity, competition, and the cost of capital again determine outcomes, and differences in management quality drive returns.
IT was the largest source of absolute return in the quarter. The portfolio's IT weight is less than one-third of MSCI ACWI's, but thanks to long-held memory-semiconductor positions, the IT segment generated an absolute return of roughly 100%, versus only 39% for the index. At the same time, the underweight made IT a modest relative drag.
Position moves: In Q2, the author continued to take profits on the DRAM memory-semiconductor basket—share prices reached record highs—realizing about 3% of profits from this long-term theme this year. For specific stocks, the author is "cautiously maintaining exposure" because these companies are essential to the AI arms race and near-term capacity expansion remains constrained. However, as a capital-cycle investor, the author explicitly warns that in a demand-driven super-cycle, record profits will eventually incentivize new supply, with unavoidable consequences for returns and share prices (see Luke Bridgeman's Hosking Post, "Super Cycle or Capital Cycle"). The table below lists the Q2 performance of the individual stocks in the basket as stated in parentheses in the original (in USD); the direction refers to trimming the basket as a whole, as the original does not specify the proportion for each stock:
| Position | Direction | Core Logic (One Line) | Key Data |
|---|---|---|---|
| Micron | Trimmed (profit-taking, exposure retained) | Memory upcycle + AI necessity; shares at record highs | Q2 +242% |
| SK Hynix | Trimmed (profit-taking, exposure retained) | Same as above | Q2 +225% |
| Seagate | Trimmed (profit-taking, exposure retained) | Same as above | Q2 +147% |
| Corning | Trimmed (profit-taking, exposure retained) | Same as above | Q2 +88% |
| TSMC | Trimmed (profit-taking, exposure retained) | Same as above | Q2 +38% |
| LAM Research | Trimmed (profit-taking, exposure retained) | Same as above | Q2 +103% |
| Applied Materials | Trimmed (profit-taking, exposure retained) | Same as above | Q2 +112% |
| BE Semiconductor | Trimmed (profit-taking, exposure retained) | Same as above | Q2 +60% |
An energy weight of roughly 11% (versus about 4% for the benchmark) was the largest absolute and relative drag of the quarter, with Brent falling from around $100 to $70; materials were the second-largest drag. The author used the PGM pullback to rotate and add to Sibanye Stillwater.
Energy (largest drag): Fragile ceasefires and the U.S.–Iran 14-point "framework agreement" caused oil prices to drop sharply. The author made opportunistic, marginal position adjustments during the period, but holdings in North American producers and offshore drillers stayed essentially unchanged—geopolitical events (the Iran conflict and the 2022 Russia–Ukraine war) did not trigger a supply response, major producers did not chase prices higher and implicitly assumed the rally would not last, and the system has little spare capacity. With inventories drawn down by conflicts, energy markets are likely to remain highly volatile, which is an opportunity for long-term investors focused on the supply side.
Sibanye Stillwater thesis: Bullish. Eighteen months ago, leverage was at troubling levels; now, with strong cash flow at high PGM prices, a net-cash balance sheet is within sight. New CEO Richard Stewart has pledged capital discipline, and the author expects shareholder returns to be prioritized, seeing scope for a substantial re-rating from the current roughly 1x normalized EV/EBITDA valuation. Risks: the original implies that a pullback in PGM prices would again hit cash flow, and the re-rating depends on management delivering on its commitments.
Position moves:
| Position | Direction | Core Logic (One Line) | Key Data |
|---|---|---|---|
| North American producers and offshore drillers (not named) | Hold/observe (opportunistic adjustments) | No supply response, depleted inventories; high volatility is the opportunity | ~11% weight vs. ~4% benchmark |
| Sibanye Stillwater | Added | Net cash in sight + capital discipline + low-valuation re-rating | c1x normalized EV/EBITDA |
| Higher-valued PGM peers (not named) | Trimmed | To fund the Sibanye addition | Not provided |
Japan's weight was raised another 1 percentage point to 13%, and Hikari Tsushin—the portfolio's largest Japan holding—was added to. The author also significantly increased the New York REITs SL Green and Vornado, betting that scarce core office space will push rents higher.
The logic for adding to Japan is deep value plus benefiting from corporate-governance reform. Hikari Tsushin is held by five multi-counsellors, and its share price has fallen back to the level when the position was first established in late 2022; its culture is focused on capital efficiency, and book value has compounded at 17% over 15 years (as detailed by Omar Malik at the MOI Global Asian Investing Summit). On New York REITs, a construction cycle spanning more than a decade was interrupted by COVID, making core office space increasingly scarce and putting upward pressure on rents.
Position moves:
| Position | Direction | Core Logic (One Line) | Key Data |
|---|---|---|---|
| Hikari Tsushin | Added | Capital-efficiency culture; valuation back to initial-build level | Book value 15-yr CAGR 17% |
| Tokyo Steel | Added | Deep value benefiting from governance reform | Not provided |
| Iwatani | Added | Same as above | Not provided |
| Asahi Group | Added | Same as above | Not provided |
| Ezaki Glico | Added | Same as above | Not provided |
| Tosei | Added | Same as above | Not provided |
| Kyocera | Trimmed (profit-taking) | Long-term holding, strong performance | Not provided |
| Mitsubishi | Trimmed (profit-taking) | Long-term holding, strong performance | Not provided |
| SL Green | Added | Scarce core office space; upward rent pressure | Not provided |
| Vornado | Added | Same as above | Not provided |
The author cites the introduction to Edward Chancellor's Capital Account to warn that heavy equity issuance by overheated industries is a precursor to deteriorating future returns on capital. The IT sector will issue substantial amounts of stock over the next 12 months. AI buildouts have consumed the free cash flow of hyperscalers, and the buyback era is nearing its end. This stock supply must be absorbed at a time when household equity allocations are at record highs—unfavorable for an increasingly concentrated index, but favorable for the strategy's contrarian structure.
The author quotes Edward Chancellor's introduction to Capital Account: "high rates of earnings retention and capital expenditure by companies leads to excess capacity, causing future returns on capital (and stock prices) to decline." In other words: "high corporate earnings retention and high capital expenditure lead to excess capacity, causing future returns on capital (and stock prices) to decline." The author believes the current situation closely resembles what the book describes: the largest sector, IT, will issue large amounts of stock over the next 12 months—a sign of elevated valuations and intensifying competition that usually foretells poor future returns on capital for that sector.
Specific signals: since the global financial crisis, U.S. companies have relied on cheap debt to buy back shares, shrink equity, and support returns and valuations. That era is nearing its end—AI buildouts have consumed the free cash flow of all hyperscalers, leaving no capacity for buybacks; Alphabet and Meta have instead conducted secondary offerings to avoid falling behind in the spending race; SpaceX's listing raised $75bn, more than double the previous record of $29bn set by Saudi Aramco in 2019; massive IPOs from large-model giants such as OpenAI and Anthropic will follow, and investment banks will encourage a wave of private companies to seize the window. J.P. Morgan expects U.S. net equity issuance to reach $1.2bn in 2027 (the original's figure; the chart's units are USD bn, with data spanning 1999–2027 and 2026/27 estimates), the largest increase in supply since 1999—at a time when U.S. household equity allocations are already at all-time highs.
Strategy comparison: the strategy is significantly underweight sectors with record capital expenditures and continuous equity issuance, and overweight industries with disciplined capital spending, rising free cash flow, and shrinking share counts. Stance [cautious]: the author lacks confidence in the future returns of an increasingly concentrated index but is confident in the strategy's own structure—note that the author uses a classic text to underwrite his contrarian positioning; this is part of the narrative argument.