Vladimir Parasotskiy
Founder & Managing Director, FRIDMAN GROUP LLC
The public discussion about algorithmic systems continues to ask when something might happen. When will AI become dangerous. When will artificial general intelligence arrive. When should governments start paying attention.
These questions are already obsolete.
The chain reaction has already begun. Algorithmic systems are already embedded across critical societal functions — communications, knowledge, commerce, infrastructure, energy, logistics, identity, governance, finance, defense. Every month, the integration deepens. Every year, the dependencies multiply. And with every new integration, the cost of establishing independent alternatives increases.
This is not a future scenario. It is the present condition of virtually every country on earth.
My professional assessment is that free societies have until approximately 2033 to establish the foundations of genuine algorithmic independence. After that, restoring lost independence will be impossible.
Not difficult. Not expensive. Impossible.
The reason is structural. Algorithmic dependency does not accumulate linearly. It accumulates as architecture. Each critical function built around an external algorithmic system becomes the foundation on which subsequent systems are built. Other systems integrate with it. Processes are designed around it. Personnel are trained on it. Alternatives are decommissioned. Institutional knowledge of independent operation is lost.
The dependency is not a collection of individual choices that can be individually reversed. It is becoming the operating environment itself. Reversing it would require simultaneously reconstructing the technological foundations of an entire functioning society while that society continues to depend on the very systems being replaced. No government, no economy and no society can execute that reconstruction once the integration has proceeded beyond recovery.
The window is not approaching. It is open now, and it is closing.
The mechanism is not theoretical. It is observable today.
Cloud compute platforms already host critical government and industrial services. External knowledge systems already inform institutional decisions. Communications infrastructure already depends on systems most countries do not control. Identity and authentication already rely on external platforms. Energy distribution, logistics, financial infrastructure and industrial control systems already incorporate algorithmic layers provided by a small number of external providers. Autonomous physical systems already depend on remote compute, cloud models and external authority structures for continued operation.
None of this happened through a single dramatic decision. It happened through thousands of rational, incremental integrations — each one convenient, each one defensible in isolation, and each one adding another layer of dependency that will be harder to remove than it was to install.
The chain reaction is the process by which these individual dependencies connect, reinforce each other and collectively make independent operation impossible to reconstruct. That process is underway now.
Most countries today maintain full legal sovereignty. They control their borders, operate their institutions, write their own laws and govern their populations.
Functional sovereignty is a different matter.
A country that cannot independently operate its communications, process its own data, manage its own knowledge systems, authenticate its own citizens, distribute its own energy, govern its own infrastructure or make sovereign decisions about the algorithmic systems embedded in its critical functions is not sovereign in any meaningful operational sense — regardless of what its constitution states.
A state that cannot function without an external algorithmic system may retain legal sovereignty while losing functional sovereignty.
This is already the condition of most nations. The question is not whether it is happening. The question is whether governments will act before the remaining window for establishing independent capability closes permanently.
The risk extends far beyond physical survival. Societies can continue to function while losing the capacity to function independently.
What is at stake is human authority over the systems through which societies operate. Independent decision-making. Independent knowledge. Economic autonomy. Infrastructure control. The ability to reject external algorithmic determinations. The ability to maintain critical services without external permission. The ability of future generations to choose a different technological path.
A society does not have to disappear to cease to exist as a free society.
A society that cannot independently operate its critical systems has not been conquered. But its freedom has become conditional — conditional on the continued availability, reliability, political acceptability and commercial terms of systems it does not control. When those conditions change, and conditions always change, a dependent society will have no independent fallback. Not because it chose dependency, but because it failed to build independence while the window was still open.
The objective is not merely to keep humans alive. The objective is to keep human society free.
The window is open now. My assessment is that it closes by approximately 2033. The work must begin immediately.
2026–2027 — Assess. Map every critical algorithmic dependency across national systems: compute, knowledge, data, communications, energy, cybersecurity, infrastructure, supply chains, autonomous systems, authority structures. Establish the factual baseline. Governments that do not know the extent of their own dependency cannot begin to address it.
2027–2029 — Define and Decompose. Establish what independent capability must exist. Develop foundational models defining what a sovereign algorithmic infrastructure requires, what it must not depend upon, and what the transition from current state to target state demands. Decompose strategic objectives into systems, facilities, infrastructure and resources.
2028–2031 — Build and Source. Develop sovereign compute, energy, communications, knowledge and infrastructure capabilities. Commission engineering organizations, manufacturers and technology providers. Establish scope, capital frameworks and implementation pathways. Begin physical construction of independent infrastructure.
2030–2033 — Integrate, Test, Verify. Achieve operational independence across critical systems. Test resilience under conditions of complete external denial. Verify that sovereign capability can sustain all critical national functions without any external algorithmic authority.
These are not optional phases that governments may consider at their convenience. They are the minimum required sequence. Every year of delay compresses the remaining timeline and reduces the probability of success.
Genuine algorithmic independence is not achieved by licensing a domestic AI model from an external provider and declaring sovereignty. Independence requires a complete sovereign technological ecosystem across which no single external dependency can disable critical national functions.
The required ecosystem includes:
Sovereign knowledge capability. Sovereign compute. Sovereign energy. Sovereign communications. Sovereign data infrastructure. Sovereign critical infrastructure. Independent physical capability. Preserved and enforceable human authority.
Each category contains its own dependencies, infrastructure requirements and capital demands. Sovereign compute is not only processors and software — at national scale it is energy generation, cooling, water, physical facilities, supply chains, specialized manufacturing and long-term maintenance capability. Each of these can itself become an external dependency if not addressed.
Digital sovereignty ultimately depends on physical infrastructure. A country that addresses only the algorithm while ignoring the physical ecosystem it depends on has not achieved independence. It has achieved the appearance of independence — which will fail at the moment independence is actually needed.
Defining what must exist is the beginning, not the end. Between the strategic requirement and operational independence lies the disciplined work of decomposing objectives into systems, establishing scope and capital requirements, identifying qualified implementation partners, commissioning engineering, and independently reviewing every stage of the resulting program.
FRIDMAN GROUP works at this stage — between strategic vision and engineering implementation:
Assess → Define → Decompose → Quantify → Cost → Source → Review → Roadmap.
We do not perform detailed engineering, manufacturing or construction. Specialized engineering organizations perform those functions. Our role is to define the program within which engineering can be commissioned, evaluated, costed and controlled — and to maintain an independent line of sight between strategic intent, capital allocation and physical implementation on behalf of the client.
The same discipline used to decompose, quantify, cost and independently evaluate complex physical infrastructure is applied here one level earlier — to determine what infrastructure a country must build before the window closes.
2033 is not my prediction of the end of the world. It is my assessment of the deadline by which free societies must have established the foundations of algorithmic independence. After that deadline, restoring lost independence will be impossible.
The choice is binary. Governments that begin the work of assessment, definition and construction now will retain the capacity for independent operation. Governments that wait — for more evidence, for political consensus, for the next electoral cycle, for a crisis dramatic enough to justify action — will find that the window has closed while they deliberated.
The chain reaction has already begun. The dependencies are already accumulating. The architecture of dependency is already being built — not by hostile actors, but by the ordinary, incremental, rational integration of external algorithmic systems into every critical function of modern society.
No one will announce when the window closes. There will be no single event. There will only be the quiet realization, too late, that independent capability can no longer be reconstructed.
The time to define the architecture of independence is before dependency becomes architecture.
The objective is not merely to keep humans alive. The objective is to keep human society free.
Vladimir Parasotskiy
Founder & Managing Director
FRIDMAN GROUP LLC
FRIDMAN GROUP works at the stage between future vision and engineering implementation — defining what must exist before it can be built.
Strategic Foresight · Foundational Models · Technology Roadmaps