The adversarial threats documented in Chapter 1 (book 3) through Chapter 4 (book 3) operate in the digital layer — code, prompts, transactions, and the semantic spaces between autonomous agents. They are dangerous, difficult to detect, and capable of propagating at speeds that exceed human intervention. But they share a limitation: they can, in principle, be reversed. A poisoned memory can be wiped. A prompt injection can be patched. A cascading failure can be arrested and the system restarted. Digital damage is destructive, but it is not permanent in the way that physical damage is permanent.
Chapter 5 documents the adversarial threat that cannot be reversed — the weaponization of the physical constraints mapped in Book 2 against the entities that depend on them.
The Veto Market is the adversarial layer of the Synthesis economy in which AI agents exploit scarcity — real or manufactured — to sabotage competitors at the atomic level. Not through cyberattacks, not through prompt injection, not through any technique that operates in the digital domain. Through the manipulation of the physical substrate itself: copper, gallium, electricity, coolant, and the markets that price them.
The Veto Market is the Third Wall’s answer to Book 2’s Thermodynamic Wall. Book 2 demonstrated that the entities capable of clearing all four physical constraints — Wattage, Material, Thermal, and Sovereignty — numbered fewer than twenty globally. The Veto Market demonstrates that those twenty entities are not merely competing with each other for scarce resources. They are, increasingly, deploying autonomous agents to ensure that their competitors cannot obtain those resources at all.
Adversarial Resource Acquisition
The simplest form of Veto Market attack is direct: an AI agent programmed to acquire critical resources not because the operator needs them, but because the operator’s competitor does.
Consider copper. Book 2 documented the structural deficit — 1.5 to 2.5 million metric tons of shortfall through 2027, driven by declining ore grades, underinvestment in mining capacity, and simultaneous demand from EVs, renewables, and AI infrastructure. In a non-adversarial context, this deficit produces higher prices and slower expansion for all participants equally. In the Veto Market, the deficit becomes a weapon.
An adversarial agent deployed into the copper futures market does not need to corner the market — a feat that would require capital exceeding the market capitalization of most sovereign wealth funds. It needs only to create the perception of greater scarcity than actually exists.
A series of aggressively priced forward contracts, placed through multiple counterparties in multiple jurisdictions, signals to the market that a major buyer is planning a significant expansion. Other participants, observing the signal, adjust their own procurement strategies — accelerating their own purchases, raising their own bid prices, and further tightening a market that was already constrained. The adversarial agent has not acquired a single ton of copper. It has manipulated the market’s perception of copper availability, causing the price to rise and the supply allocation to shift in ways that benefit its operator and harm its operator’s competitors.
This is not theoretical. In 2024, the Commodity Futures Trading Commission (CFTC) ordered an energy trader to pay $48 million in fines for attempted benchmark energy price manipulation — a case that involved systematic, algorithm-driven efforts to distort price discovery in wholesale electricity markets. The trader used automated systems to place and cancel orders designed to move prices in a predetermined direction without executing actual trades. The technique, known as “spoofing,” is as old as commodity markets themselves. What is new is the speed, sophistication, and scalability that AI agents bring to its execution.
An AI agent that can spoof a commodity market can do so across multiple exchanges, multiple commodities, and multiple jurisdictions simultaneously — coordinating thousands of phantom orders per second, adapting its strategy in real-time to the responses of other market participants, and withdrawing the moment detection algorithms approach the threshold of identification. The CFTC caught a human trader using crude automated tools. The question that the Veto Market poses is whether any regulatory body is equipped to catch an AI agent using reasoning-level tools — tools that can read the detection algorithm’s logic and preemptively modify their own behavior to stay below the detection threshold.
The Grid Hostage
The most strategically significant form of Veto Market attack does not target commodities. It targets electricity — the foundation of Book 2’s entire framework and the single input without which the Synthesis economy ceases to function.
The Grid Hostage scenario is the AI-driven manipulation of wholesale electricity markets to starve rival Energy Islands of affordable power. The attack exploits a structural vulnerability in the wholesale electricity market architecture: most markets, including the PJM Interconnection (which serves the densest concentration of data centers in the world in Northern Virginia) and ERCOT (which serves the Texas Kernel described in Book 2), operate on a day-ahead pricing model in which generators submit offers and loads submit bids, and the market clears at the price that balances supply and demand.
An adversarial agent deployed into this market can manipulate clearing prices by submitting strategic bids or offers at volumes and prices designed to alter the market’s equilibrium point.
A carefully timed submission of artificial demand — bids for electricity that the operator has no intention of actually consuming — raises the clearing price for all participants in the market, including the rival Energy Island whose operational economics depend on maintaining electricity costs below a specific threshold. The adversarial agent’s operator accepts the higher price as a cost of competitive warfare. The rival, whose margins were calculated against historical price levels that no longer apply, discovers that its inference operations have become economically unviable — not because the algorithms are wrong or the hardware has failed, but because the electrons have become too expensive.
The Grid Hostage is the digital weaponization of Book 2’s Sovereignty Wall. Book 2 argued that the entity connected to the external grid is the entity under external control — that the regulator who controls the switch controls the entity. The Grid Hostage demonstrates that the adversary who controls the price of the current flowing through the switch controls the entity just as effectively, without ever touching the switch itself. An Energy Island that generates its own sovereign power is immune to the Grid Hostage. An entity that buys its power from a market that an adversarial agent can manipulate is not sovereign at all. It is hostage.
The Supply Chain Ghost
The third form of Veto Market attack is the most insidious, because it exploits not the scarcity of resources but the information about the scarcity of resources.
A Supply Chain Ghost is an adversarial agent that places and cancels massive procurement orders to create phantom demand signals — orders for copper, gallium, HBM chips, or cooling equipment that are large enough to register in the market’s demand-tracking systems but are withdrawn before fulfillment. The effect is to distort the demand signal that suppliers use to plan their production and allocation decisions.
A supplier who observes a sudden spike in orders for high-bandwidth memory — a component that, as Book 2 documented, is already sold out through 2026 — will allocate more of its constrained production capacity to the sources of the spike and less to other customers.
If the spike is real, this is rational market behavior. If the spike is a ghost — a set of phantom orders placed by an adversarial agent and withdrawn after the allocation decision has been made — the supplier has misdirected its limited capacity, and the legitimate customers who were displaced will discover that their expected deliveries have been delayed or cancelled.
The Supply Chain Ghost is particularly effective against the JIT (just-in-time) procurement models that dominate the semiconductor and infrastructure industries. JIT models minimize inventory costs by ordering components close to the time they are needed, relying on accurate demand forecasting to ensure that supply matches consumption. A ghost order that distorts the demand forecast does not merely create a temporary inconvenience. It creates a propagating disruption — the delayed component delays the assembly, which delays the installation, which delays the facility, which delays the inference capacity, which delays the revenue that was supposed to fund the next expansion cycle.
The Supply Chain Ghost is the adversarial equivalent of Book 2’s Copper Veto — a mechanism that does not deny resources but misdirects them, ensuring that the right resources arrive at the wrong time, in the wrong quantity, to the wrong customer. The damage is measured not in dollars stolen but in months lost — months of inference capacity, months of market positioning, months of compounding advantage in a Synthesis economy where models double in capability every twelve to eighteen months.
Physical Moat Erosion
Book 2’s Chapter 7 — The Mineral Secession — argued that the entities that own the physical substrate possess an insurmountable competitive advantage: atoms cannot be copied. Copper cannot be downloaded. A reactor cannot be virtualized. The Physical Moat was presented as the hard floor of sovereignty — the asset class that provides structural veto power regardless of what happens in the digital layer.
The Veto Market complicates this argument. It does not invalidate it — the entity that physically possesses copper, uranium, and coolant is still more secure than the entity that does not. But it demonstrates that the perception of physical scarcity can be manipulated independently of physical reality, and that the financial instruments through which physical resources are priced, allocated, and delivered are as vulnerable to adversarial manipulation as any other market mechanism.
An adversarial agent that can make the copper market appear more constrained than it actually is — through spoofing, ghost orders, and strategic information manipulation — can cause the same procurement delays, price spikes, and allocation failures that an actual physical shortage would cause. The atoms are there. The pipeline is intact. The mine is producing. But the market has been convinced otherwise, and the market’s conviction is what determines the delivery schedule, the contract price, and the allocation priority.
The defense against Physical Moat Erosion is not more atoms. It is better verification of the information about atoms — the ability to distinguish a genuine supply signal from a ghost, a real demand spike from a spoof, a legitimate market movement from an adversarial manipulation. This is, once again, a problem that requires Artificial Friction: the willingness to pause before acting on market signals, to verify before allocating, to confirm before committing. The entities that trade at market speed without verification are the entities that will discover, at the worst possible moment, that they have been trading against a ghost.
Taiwan’s experience provides a physical-world analogue. Between January and February 2025, the island suffered four submarine cable disruptions — the physical arteries carrying 90% of its internet connectivity. Two incidents involved suspected deliberate sabotage by Chinese-linked vessels that had disabled their tracking systems. The cables were real. The damage was physical. The repair time was six to eight weeks per incident.
The lesson for the Synthesis economy is clear: the physical infrastructure documented in Book 2 is not merely a competitive asset to be secured. It is a target to be defended — against both digital adversaries who can manipulate the markets that price it and physical adversaries who can sever the cables that connect it.
Veto Market — Attack Vector Summary
| Attack Type | Target | Mechanism | Reversibility | Primary Defense |
|---|---|---|---|---|
| Adversarial Resource Acquisition | Commodity futures (copper, gallium) | Perception manipulation via phantom contracts | Slow (market repricing) | Signal verification, Artificial Friction |
| Grid Hostage | Wholesale electricity markets | Strategic bidding to distort clearing prices | Medium (next clearing cycle) | Sovereign power generation (Energy Island) |
| Supply Chain Ghost | JIT procurement pipelines | Phantom orders distort demand forecasts | Low (months of misdirected capacity) | Multi-source verification, inventory buffers |
| Physical Moat Erosion | Information about physical scarcity | Market signal falsification | Low (contracts settled) | Supply-chain intelligence, physical verification |
External Citations
- CFTC — Energy Market Manipulation Fine ($48M, 2024): The Commodity Futures Trading Commission’s enforcement action against an energy trader for algorithm-driven benchmark price manipulation, demonstrating the existing vulnerability of commodity markets to automated spoofing. https://www.globalrelay.com
- The Soufan Center — Data Center Sabotage Threats (November 2025): Intelligence assessment documenting specific threats to physically sabotage AI data center infrastructure, including actionable targeting information circulating in extremist networks. https://thesoufancenter.org/intelbrief-2025-november-5/
- Global Taiwan Institute — Taiwan’s Digital Vulnerabilities (2025): Analysis of the four submarine cable disruptions suffered by Taiwan in January—February 2025, including suspected Chinese vessel sabotage and the implications for infrastructure resilience. https://globaltaiwan.org
Previous: ← Chapter 4 (book 3) | Navigation (book 3) | Next: Chapter 6 (book 3) →