Daniil Liberman is sounding the alarm about who controls the backbone of artificial intelligence. In a recent analysis, he argues that the concentration of AI infrastructure in the hands of a few threatens individual autonomy and that equitable access is essential for preserving freedom. The stakes, he says, are as high as those seen in past financial crises.
Why AI infrastructure control matters
At the heart of Liberman's argument is a simple but stark claim: whoever controls the computing power, data pipelines, and algorithmic gateways of AI holds enormous power over how the technology is used — and for whose benefit. Without broad access, he warns, the very tools that could expand human potential may instead become instruments of control. Liberman draws a direct line between this dynamic and historical battles between centralization and decentralization, which he sees playing out again in today's tech landscape.
It's not just about who builds the next large language model or runs the biggest data center. Liberman contends that the architecture of AI — its hardware, software, and governance — will shape political and economic freedom for decades. If a small number of companies or governments lock down that infrastructure, he argues, ordinary people lose the ability to shape their own digital lives.
Lessons from past financial crises
Liberman doesn't stop at abstract theory. He compares the risks of AI centralization to the systemic failures that triggered past financial crises. In those episodes, concentrated power and opaque decision-making led to collapses that wiped out savings and deepened inequality. The same pattern, he suggests, could repeat with AI — only faster and more broadly. The demand for decentralized solutions, he notes, is already rising as a response. People are looking for alternatives that distribute control rather than concentrate it.
The historical parallel isn't accidental. Liberman sees a recurring cycle: innovation leads to concentration, concentration creates vulnerability, and vulnerability sparks a push for decentralization. AI, in his view, is the latest frontier of that cycle.
The push for decentralized AI
What does a decentralized AI infrastructure look like? Liberman points to models that spread computational resources across many participants, open-source algorithms that anyone can audit, and governance structures that include a wide range of voices. These aren't pipe dreams. Projects exploring decentralized computing, community-owned data pools, and cooperative AI development are already underway. But they face steep hurdles: funding, coordination, and the sheer inertia of today's centralized giants.
Liberman's argument is not a call to tear down existing systems but to build viable alternatives. He insists that equitable access isn't a luxury — it's a prerequisite for any future where AI serves human freedom rather than undermining it. Without that access, he warns, the technology's benefits will flow only to the already powerful.
The question now is whether policymakers, technologists, and the public will take that warning seriously before the next infrastructure crisis hits.
