Program: Global System & Power Code: MT-SA-2026-09-19-structural-power Edition: September 19, 2026 Information cutoff: September 19, 2026
Large pools of wealth matter, but wealth alone does not explain structural power. An asset manager can direct client capital without owning it. A commodity trader can sit inside a physical supply chain without owning most of the resources it moves. A technology company can have a much smaller balance sheet than a global fund and still occupy a difficult-to-replace position in a production system. A government can regulate access to that technology without operating the company that produces it.
The analytical problem is therefore not to compile a list of rich institutions. It is to identify which actors can change the constraints under which other actors operate, through which documented mechanism, for how long, and with what substitutes available.
This first Strategic Actors assessment establishes that method and tests it against five different structures: BlackRock, Norges Bank Investment Management / Government Pension Fund Global, Trafigura, Glencore and ASML. The comparison is intentionally heterogeneous. The purpose is to show why assets under management, shareholder rights, physical intermediation, productive control and technological dependence cannot be collapsed into one ranking.
The central assessment is: structural capacity is best understood as a chain linking an actor to a capability, a dependency and a transmission mechanism. Size matters when it strengthens that chain, but size is not the chain itself.
The first error is treating assets under management as ownership
BlackRock reported US$15.3 trillion in assets under management at June 30, 2026, after US$868 billion of net inflows over the previous twelve months. That number describes the scale of assets managed for clients. It does not mean BlackRock owns US$15.3 trillion of securities for its own economic account. BlackRock — Q2 2026 results
The distinction matters because BlackRock does have a corporate-governance transmission channel, but it is a different one. BlackRock Investment Stewardship states that it votes at shareholder meetings for clients who have authorized it to vote on their behalf and engages with boards and management to inform those voting decisions. BlackRock also states that more than 90% of its clients' public-equity AUM was invested in index equity strategies at June 30, 2026. BlackRock Investment Stewardship
The defensible chain is therefore:
- Client savings and institutional capital
- investment mandate
- BlackRock portfolio implementation
- securities held for clients
- delegated voting and engagement where authorized
- corporate-governance channel
The final arrow is real, but bounded. Voting authority depends on the mandate and client authorization. Engagement does not establish operational control. Index ownership also constrains the ability to exit individual companies in the same way as an unconstrained active owner. A serious map of power must preserve those limits rather than replace them with the word "controls."
Five actors, five different mechanisms
| Actor | Documented structural channel | Observed scale | What the evidence does not establish |
|---|---|---|---|
| BlackRock | delegated capital allocation and shareholder voting | US$15.3tn AUM at 30 Jun 2026 | ownership of client assets or unilateral control of portfolio companies |
| NBIM / GPFG | public-mandated global ownership and shareholder rights | NOK21.268tn fund value at end-2025 | direct control of thousands of portfolio companies |
| Trafigura | physical commodity intermediation, logistics and downstream links | 6.6m barrels of oil and petroleum products traded per day in FY2025 | ownership of global oil reserves or unilateral price control |
| Glencore | owned/controlled production plus physical marketing | 851.6kt copper, 969.4kt zinc and 98.0Mt energy coal own-source production in 2025 | unilateral control of multi-producer global commodity markets |
| ASML | specialized lithography technology for semiconductor manufacturing | 535 systems sold in 2025, including 48 EUV systems | control of the entire semiconductor stack or sovereign authority over export licences |
These quantities are not comparable on a common axis and should not be plotted as if they were. The table is a map of mechanisms, not a scale ranking. Sources: NBIM Annual Report 2025, Trafigura Annual Report 2025, Glencore FY2025 Production Report, ASML 2025 Annual Report.
Physical intermediation can matter without a trillion-dollar balance sheet
Trafigura reported total oil and petroleum-products trading volume of 318.2 million metric tonnes in FY2025 and an average 6.6 million barrels traded per day, compared with 6.0 million in FY2024. The company also operates through a network that includes downstream businesses and storage infrastructure; Puma Energy, for example, describes a network of more than 80 storage terminals, while TFG Marine supplies marine fuel through strategic hubs. Trafigura Annual Report 2025 Trafigura in brief
View data
| Indicator / period | Value (million barrels per day) |
|---|---|
| FY2024 | 6 |
| FY2025 | 6.6 |
The chart measures physical trading volume, not reserves owned, market share or price-setting power. Its analytical value is different: it shows that an actor can be structurally relevant because it is repeatedly present in the matching, financing, storage and movement of physical supply.
That channel becomes more important when markets are disrupted. A trader with financing capacity, shipping access, storage, counterparties and operational information may be able to reroute flows that a passive asset owner cannot. Yet that does not imply unilateral control. Commodity trading remains competitive, contracts can expire, logistics can be replaced, and producers and consumers can use other intermediaries.
Glencore represents a different configuration because physical intermediation is combined with owned or controlled industrial assets. In 2025 it reported own-source production of 851.6 thousand tonnes of copper, 969.4 thousand tonnes of zinc, 71.9 thousand tonnes of nickel, 36.1 thousand tonnes of cobalt, 98.0 million tonnes of energy coal and 32.5 million tonnes of steelmaking coal. It also operates a separate marketing business. Glencore FY2025 Production Report
The mechanism is therefore closer to:
- Owned or controlled mines and industrial assets
- physical production
- marketing and trading network
- processors and end users
- prices, inventories and industrial availability
Again, the last stages depend on market conditions and competing suppliers. Production control is stronger than pure intermediation, but it is still not equivalent to control of the entire commodity market.
Technological dependence changes the meaning of scale
ASML illustrates why financial size can be the wrong starting point. In 2025, the company reported €32.7 billion of net sales, €4.7 billion of R&D and 535 systems sold, including 48 extreme-ultraviolet lithography systems and 279 deep-ultraviolet systems. ASML 2025 Annual Report
The important question is not whether those figures are larger or smaller than a sovereign fund's portfolio. The question is whether downstream production can substitute the capability.
The Dutch government explicitly subjects categories of advanced semiconductor manufacturing equipment, including lithography equipment, to national export authorization requirements. In January 2025 it expanded those controls and stated that the Netherlands has a unique role in semiconductor manufacturing technology; licensing remains a case-by-case government decision rather than an export ban. Government of the Netherlands — advanced semiconductor equipment export controls
That distinction creates two separate actors in the same structural chain:
- ASML R&D and manufacturing
- advanced lithography systems
- semiconductor fabrication capability
- advanced chips
- computing, industrial and security applications
- Dutch state export-control framework
- authorization or restriction of specified equipment exports
- geographic access to advanced manufacturing capability
ASML supplies the technology. The Dutch state exercises sovereign licensing authority. Conflating those roles would be as misleading as conflating BlackRock's client AUM with BlackRock's own balance sheet.
Substitutability is the variable that turns capacity into dependency
A large actor can be replaceable. A smaller actor can occupy a hard-to-replace point. Structural analysis therefore needs to ask not only how much an actor owns or moves, but how quickly another actor can reproduce the function.
Capital
- allocation mandates
- credit and liquidity
- sensitivity to long-term interest rates of capital
Ownership and governance
- beneficial ownership
- voting authority
- board or contractual rights
Physical system
- production
- storage
- transport
- processing
Technology
- proprietary capability
- manufacturing know-how
- standards and interfaces
State linkage
- public mandate
- licensing
- regulation
- procurement
Dependency
- substitutability
- switching time
- switching cost
- geographic concentration
Persistence
- recurring flows
- durable assets
- institutional mandate
This framework deliberately avoids a single "power score." Combining US dollars of AUM, barrels per day, voting rights and technology substitution into one number would manufacture precision rather than measure a coherent variable.
The more useful output is a profile: which channels are documented, what they reach, what limits them, and whether the dependent system has credible alternatives.
Public funds add another layer: the mandate must be separated from the portfolio
The Government Pension Fund Global ended 2025 at NOK21.268 trillion, with 71.3% in equities, 26.5% in fixed income, 1.7% in unlisted real estate and 0.4% in unlisted renewable-energy infrastructure. Its governance chain is explicit: Norway's parliament establishes the formal framework, the Ministry of Finance has overall responsibility and issues management guidelines, Norges Bank manages the fund, and NBIM performs operational management. NBIM Annual Report 2025 NBIM governance structure
That architecture matters because "state money" does not describe how decisions are actually made. The same is true across sovereign investors. Some have explicit domestic transformation mandates; others are designed primarily to preserve or compound public wealth; others own strategic companies while maintaining commercial decision-making structures.
The second research in this initial release therefore examines sovereign capital separately rather than assuming a single model.
What the first map changes
The first Strategic Actors dataset treats every relation as a typed mechanism. It distinguishes:
| Relationship | What it means | What it must not be silently upgraded into |
|---|---|---|
| delegated capital allocation | an actor implements client or beneficiary mandates | ownership of the underlying wealth |
| delegated voting | an actor votes where authorized | direct operational control |
| direct ownership | an actor owns the asset or security | automatic control when rights are minority or dispersed |
| state mandate | a public institution has an explicit mandate | proof that every individual transaction is geopolitical |
| commercial intermediation | an actor moves or matches material flows | ownership of the underlying resource base |
| production control | an actor owns or controls productive assets | control of the entire global market |
| technology supply | an actor supplies a difficult-to-replace capability | sovereign authority over who may receive it |
The public tool uses those distinctions to connect actors to markets, industrial capacity, energy flows, critical materials and advanced semiconductor production. The database is intentionally selective. Absence from the map means only that an actor has not yet passed the evidence and materiality test.
What should enter next
The next expansion should not begin with famous surnames. It should begin with mechanisms that remain under-mapped.
That means private-credit platforms that have become material lenders to infrastructure and corporate borrowers; clearing and settlement institutions whose failure would interrupt financial markets; insurers whose balance sheets determine which risks can be financed; port, shipping and pipeline operators at concentrated corridors; cloud and datacenter operators where compute dependency is geographically concentrated; semiconductor-equipment and materials suppliers; policy banks; pension funds; and legally identifiable family offices or foundations where a documented channel reaches capital, institutions or public policy.
Rothschild, Rockefeller, Soros or any other family or individual should enter only through a specific legal or operational entity. The surname itself is not an analytical node.
What would weaken an actor's structural relevance
The assessment should be revised downward when one or more of the following occur: credible substitutes scale rapidly; switching costs collapse; a mandate is materially narrowed; voting or ownership rights decline; production or flow volumes lose relevance; a formerly proprietary technology becomes commoditized; regulation removes access to the relevant channel; or a previously concentrated dependency becomes distributed among many alternatives.
The opposite also matters. Consolidation, rising market dependence, expanding mandates, vertical integration, new infrastructure ownership or tighter technological scarcity can increase structural relevance even when an actor's headline asset value changes little.
That is why Strategic Actors should be updated primarily event-driven, with periodic structural reviews. The unit being monitored is not celebrity. It is the mechanism.
Principal sources
- BlackRock — Second Quarter 2026 results.
- BlackRock Investment Stewardship.
- Norges Bank Investment Management — Annual Report 2025.
- Norges Bank Investment Management — governance structure.
- Trafigura — 2025 Annual Report.
- Trafigura in brief.
- Glencore — Full Year 2025 Production Report.
- ASML — 2025 Annual Report.
- Government of the Netherlands — advanced semiconductor manufacturing equipment export controls.