From Rosalind to Mandala
How the BIS is Building a Planetary Control System
On paper, the Bank for International Settlements (BIS) Innovation Hub is merely modernising the financial system. Its projects — Rosalind, mBridge, Mandala — are pitched as solutions for faster payments and smarter compliance.
Look closer, and a different picture emerges. These are not isolated proofs of concept; they are interlocking components of a planetary-scale adaptive management system, where money becomes a programmable control surface for policy enforcement.
The individual projects form a functional hierarchy:
Micro (Retail Control): Project Rosalind1.
Developed with the Bank of England, Rosalind builds the API layer for a retail Central Bank Digital Currency (CBDC). Its two-tier model — a central bank ledger with private-sector applications on top — is not merely a digital replica of cash. It is explicitly a platform for programmable functions, turning payments into a surface where policy can be embedded as features.Cross-Border (Wholesale & Retail): mBridge2 and Icebreaker3.
Project mBridge is a multi-CBDC platform for wholesale cross-border settlement that reached Minimum Viable Product (MVP) status in 2024. Project Icebreaker explored a hub-and-spoke model for linking retail CBDCs across borders. Together, they ensure the programmable control surface does not stop at national frontiers.Macro (The Rule Engine): Project Mandala4.
This is the master policy kernel. Mandala encodes jurisdiction-specific rules — from sanctions screening to capital-flow measures — into a common protocol for cross-border payments. It uses a ‘compliance-by-design’ approach and privacy-enhancing checks to automate regulation. In essence, it is policy-as-code for international finance.
Viewed together, the architecture is unmistakable: Rosalind is the actuator at the retail edge, mBridge and Icebreaker are the cross-border connective tissue, and Mandala is the brain that decides what transactions may execute.
The Pattern: A Live Cybernetic Loop
To understand the profound shift happening here, we need to step back from the technical jargon. What’s being built isn’t just better software — it’s a new kind of governance mechanism centered around finance, based on cybernetics. The simplest way to understand cybernetics is to think of a thermostat in your home.
A thermostat constantly senses the temperature, decides if it’s too hot or cold compared to your setting, and acts by turning the heating or cooling on or off. It’s a self-regulating system that maintains a desired state without constant human intervention.
The BIS projects collectively create precisely this kind of system, but for the entire global economy. Let’s break down this ‘planetary thermostat’ piece by piece.
Phase 1: Sense — The Planetary Nervous System
Every control system needs information. You can’t regulate what you can’t measure. The ‘Sense’ phase is about building the equivalent of a global nervous system for finance — technologies that provide real-time awareness of economic activity.
Project Pyxtrial5 acts like a financial MRI machine for stablecoins. It constantly scans and analyses the health of these digital currencies, monitoring both their on-chain liabilities and the off-chain assets that back them. This gives regulators an unprecedented real-time view into a rapidly growing part of the financial system that was previously opaque.
Project Atlas6 functions as a global financial cartographer. It maps the flow of money across borders, tracking how value moves between traditional banking systems and cryptocurrency networks. Where traditional economic data might take months to compile, Atlas provides a live, updating map of global capital movements.
Project Aurora7 creates what we might call a ‘collaborative immune system’ for financial crime. It allows banks and regulators across different jurisdictions to pool information about suspicious transactions while preserving privacy through advanced cryptography. This means that patterns of money laundering or terrorist financing that might be invisible to any single institution can now be detected by the network as a whole.
Together, these projects (and others) create a comprehensive sensory apparatus. They transform finance from a sector where understanding came from quarterly reports and delayed statistics into one where regulators have what amounts to real-time X-ray vision into the system’s inner workings.
Phase 2: Decide — The Policy Brain
Sensing alone isn’t enough. A nervous system needs a brain to interpret signals and make decisions. In this architecture, the ‘Decide’ phase is where human policy goals get translated into machine-executable rules.
Project Mandala serves as the core processing unit of this system. Think of it as a global policy compiler. It takes the complicated rulebooks of different countries — their financial regulations, sanctions lists, capital controls — and translates them into a common digital protocol. When a cross-border payment is initiated, Mandala automatically checks it against all relevant regulations simultaneously. Using advanced zero-knowledge proofs, it can even verify that a transaction complies with rules without revealing sensitive details about who’s sending what to whom.
Project Gaia8 and Project Symbiosis9 extend this decision-making capability into environmental policy. Gaia uses artificial intelligence to analyse corporate climate disclosures, automatically checking commitments against actual performance. Symbiosis focuses specifically on supply chain emissions. These projects mean that environmental, social, and governance (ESG) criteria become measurable, verifiable inputs that can feed directly into financial decision-making.
This represents a fundamental shift in how regulation works. Traditionally, rules were written in legal language that humans interpreted and enforced, often after the fact. In this new model, regulation becomes ‘policy-as-code’ — rules written in a language machines can execute automatically and in real-time.
Phase 3: Act — The Economic Muscles
Once a decision is made, it needs to be enforced. The ‘Act’ phase represents the muscles of this system — the mechanisms that physically allow or prevent economic activity based on the decisions coming from the policy brain.
Project Rosalind represents the fine motor control at the retail level. Through its API framework for CBDCs, it enables programmable money. This isn’t just digital currency — it’s currency with conditions attached. A government stimulus payment could be programmed to only be spent on essentials. A corporate subsidy could be restricted to purchases from certified green suppliers. The promise is real but the power it grants is unprecedented.
Project mBridge and Project Icebreaker provide the gross motor functions for cross-border movement. They ensure that this programmability doesn’t stop at national borders. mBridge connects wholesale CBDCs between central banks, meaning large international transactions between financial institutions inherit the same conditional logic.
The crucial thing to understand about this ‘Act’ phase is that enforcement becomes embedded, immediate, and automatic. There’s no need for a regulator to send a cease-and-desist letter or for a court to issue an injunction. Non-compliant transactions simply fail to execute.
Phase 4: Learn — The Adaptive Memory
A simple thermostat has a fixed setting. A smart thermostat learns from patterns and optimises for efficiency. The most sophisticated aspect of this emerging financial infrastructure is its capacity to learn and adapt — to become smarter over time.
Project Ellipse10 functions as the system’s analytical cortex. It takes the enormous streams of data generated by financial transactions and regulatory checks and looks for patterns, anomalies, and opportunities for optimisation. If certain compliance rules are constantly triggering false positives, Ellipse can detect this and suggest adjustments.
Project Hertha11 specialises in pattern detection within payment systems. By analysing real-time payment data across jurisdictions, it can identify emerging threats — new money laundering techniques, organised fraud rings, or systemic vulnerabilities — before they cause significant damage.
This learning capability transforms the system from static regulation to adaptive governance. The rules themselves can evolve based on their observed effectiveness. The system becomes not just a regulator but an ever-improving manager of the economic ecosystem.
From Theory to Reality: The Loop in Action
This is rapidly becoming operational reality. The BIS itself has stopped treating these as separate projects and now explicitly refers to them as components of a ‘global financial integrity technology stack’12.
Imagine this cybernetic loop operating in practice: The system senses unusual capital flows from a particular jurisdiction (Atlas), cross-references this with AML alerts (Aurora), decides this matches a risk pattern that requires intervention (Mandala), and automatically acts by imposing temporary controls on certain transaction types (through mBridge and Rosalind). Then, it learns from the outcome (using Ellipse and Hertha) to refine its detection algorithms for future events.
The entire process happens in near-real time, with minimal human intervention. This is the paradigm shift: from delayed, human-driven regulation to instantaneous, automatic, algorithmically-executed governance.
The Blueprint: Six Rails of Operational Control
This technical architecture operationalises the six rails, which implement three classic frameworks for managing complex systems.
General Systems Theory (The Network):
Digital Identity:
Digital Wallets become ‘parameterised nodes’, with credentials and permissions defining an actor’s status within the economic network.Accreditation:
This creates the trust fabric, certifying which entities and auditors are authoritative, thereby controlling which connections (‘edges’) between nodes are allowed.Together, they transform an open market into a permissioned graph.
Input-Output Analysis (The Flows):
Data:
Provides continuous, granular measurement of financial flows (Pyxtrial on stablecoins, Atlas on cross-border movements).Audit & Assurance:
Analytics like Ellipse and Aurora validate these flows against mandated patterns, flagging variances for action.This turns economic analysis into a real-time control mechanism.
Cybernetics (The Enforcement):
Finance:
This is the system’s actuator. Credit, payment permissioning, and settlement are driven by the outputs of the policy kernel (Mandala) via the rails (Rosalind, mBridge). Money becomes the tool that automatically steers behaviour.Procurement:
When public procurement mandates these standards, they cascade unavoidably through global supply chains, making adoption mandatory for market access.
The result is a recursive steering system: programmable money + machine-readable rules + permissioned identity = live adaptation of economic behaviour to meet predefined policy targets.
The Return of ‘International Government’
This technical stack resurrects a political concept from a century ago: Leonard Woolf’s vision of ‘international government’ mediated by supposedly neutral standards. Today, Mandala operationalises the undemocratic ‘regulation stack’ — the rules set by the Basel Committee (prudential), the Financial Stability Board (systemic risk), the FATF (anti-money laundering), and the NGFS (climate risk) — translating them into machine-executable constraints on transactions.
The power shift is increasingly measurable and formally acknowledged. The EU’s MiCA ‘lays down uniform requirements’ for crypto-assets and service providers across the Union, creating a single supervisory rulebook13. The BIS’s Project Mandala ‘preserves the existing regulatory framework’ by encoding jurisdiction-specific requirements — ‘such as sanctions screening and capital flow management measures’—into transactions, ie compliance-by-design14.
As Singapore’s MAS put it, Mandala shows that ‘policy compliance can be embedded in Cross-border transactions’15.
The Ultimate Enforcement: Conditional Participation and Technical Erasure
The system’s ultimate power isn’t just conditional participation — it’s practical exclusion via network rules. In 2012, after an EU Council decision, SWIFT technically disconnected sanctioned Iranian banks from its messaging network — no new treaty, just a switch flipped on the rails that route cross-border payments16. In 2022, EU regulations required SWIFT to remove designated Russian banks, producing immediate, infrastructure-level isolation from much of global finance17.
More broadly, when a country is flagged for AML/CFT weaknesses18, banks often ‘de-risk’ by cutting correspondent ties. Those relationships have dropped markedly — roughly 20% over seven years — with the steepest losses in weaker-control jurisdictions. The result is tighter access to cross-border payments even without any formal ban19.
The World Bank and the IMF work traces this to compliance pressure and reputational risk, showing how technical non-interoperability and withdrawal of correspondent lines can function as de facto exclusion20.
This power already reaches individuals and firms through the rails we have today. In February 2022, under Canada’s Emergencies Act, banks froze 180–210 accounts linked to protest activity — an immediate, infrastructure-level restriction executed via compliance orders rather than new legislation or court trials21. In the UK, the 2023 ‘debanking’ furore around Nigel Farage showed how account closures can occur through internal risk/compliance processes, prompting regulatory reviews and proposed legal safeguards on notice and reasons for closure22.
This exposes the fatal constitutional flaw: the system assumes perfect code and benevolent operators, but provides scant recourse for error.
Conclusion: The Constitutional Void in a Coded Economy
What’s emerging is not just a better financial system; it’s a planetary adaptive-management system whose feedback loop closes through conditional finance. The monetary unit of account is becoming the actuator: policy is compiled into machine-readable rules; those rules condition access to payment, credit, and settlement; behaviour adjusts to maintain target states, and the loop re-tunes itself from observed outcomes. That’s governance by programmable clearance.
The benefits are real — speed, integrity, coordination — but the constitutional risk is larger: when the unit of account itself carries policy, participation becomes contingent and error becomes economic silence.
If we accept an adaptive system, we will ultimately face a constitutional void.
The time for polite questioning is over. What’s required is a demand for:
Money rights. We have fundamental rights regarding speech, assembly, and property. We must now articulate fundamental rights regarding economic participation in a digital system. The right to a basic economic identity, the right to transact, the right to privacy.
Due process: published, versioned, machine-readable rule sets; auditable denial reasons; rapid appeals with human override.
Separation of functions: clear boundaries between rule authors, rule coders, and rule executors; mandatory transparency on procurement cascades.
Accountability: without accountability, the system creates a zone of impunity where errors become permanent economic death sentences with no responsible party to answer for the harm.
What’s particularly disturbing is how they’ve solved the ‘last mile’ problem of enforcement. Rosalind’s API layer means even private sector applications built on top of CBDCs inherit the policy constraints. You can’t escape the system by using a fintech app — the constraints are baked into the monetary unit itself. This creates what you might call ‘regulatory monoculture’ where all financial activity must conform to the same machine-readable rules.
The most dangerous fiction is that this is ‘neutral technology’. Every parameter in these systems — what constitutes a suspicious transaction, what triggers capital controls, how environmental criteria are measured — represents a political choice disguised as technical necessity. When Gaia’s AI determines your carbon footprint exceeds acceptable limits and automatically restricts your purchasing power, who do you appeal to?
The algorithm?
The operating system for the planetary economy is being deployed. Its user manual — a constitution for this new digital realm — remains unwritten.
If we fail to write it, the ‘silent architecture’ will enforce compliance not through overt coercion, but through the quiet, efficient, and inescapable elimination of non-conformant economic activity.
And every transaction will be conditional upon clearance, especially yours.
This goes into the blog post I'm composing.
Yesterday had an ESC triple...
pondering how the underground economy and black budget activity survive in this scenario. let's say the US democratic party wants to launder some mob money into their accounts stateside thru ukraine. does the BIS manually tweak the system not to flag it?
how about your local weed dealer? many communities globally are reliant on continued illegal cash flows. seems to me that crashing the underground economy would crash the one above ground as well