The Soviet AI Developed in Secret in 1965… Why It Was Shut Down

On 1 October 1970, a group of men gathered around a table in the Kremlin and voted on whether a machine should be permitted to help run the Soviet Union. The two most senior figures in the country were absent. Leonid Brezhnev was in Baku, and Alexei Kosygin was in Cairo, attending the funeral of Gamal Abdel Nasser. The mathematician who had spent eleven years designing the machine sat to one side, having been told to expect approval. By the end of that afternoon, the project that American intelligence had once described as a threat to the balance of the Cold War was effectively finished, and the man who ended it had done so by talking about chickens. To understand how a nationwide electronic brain, drafted under classification and backed by the Soviet prime minister himself, could be dismantled in a single meeting, one must go back to a letter written in 1959 by a colonel who was about to lose everything.

In 1959, the Soviet Union was attempting to run the second-largest economy on Earth with pencils, paper, and the abacus. That is a plain description, not a figure of speech. The state planning committee, known as Gosplan, was responsible for deciding how much steel, how much grain, how many tractors, how many pairs of shoes, and how many light bulbs the entire country would produce in a given year. It then had to decide where every one of those items would go. By the late 1950s, the number of distinct products tracked by Soviet planners had passed twenty million. Every one of them required a production target, a supply chain, a price, and a delivery schedule, and every one of those numbers had to be reconciled against every other number by hand. The planning cycle worked in one direction and then the other. Factories sent their reports upward to regional authorities, regional authorities compiled them and sent them to ministries, and ministries compiled them again and sent them to Gosplan in Moscow. Gosplan drafted a plan and sent targets back down the same chain. At every step, the data was copied by hand, retyped, rounded, delayed, and in many cases quietly adjusted by officials who had an incentive to report that things were going better than they were. By the time a number from a factory in Novosibirsk reached a planner in Moscow, it could be six months old and had passed through a dozen offices. The planner then made a decision based on it, and that decision took another six months to travel back down to the factory floor.

A mathematician in Kiev named Viktor Glushkov ran the numbers on this system and arrived at a conclusion he would repeat for the rest of his life. The volume of planning calculations in the Soviet economy was growing faster than the economy itself. If the country kept planning the way it was planning, then by the late 1970s it would need the labor of its entire adult population just to do the arithmetic of running itself. Nobody would be left to work in the factories the plan was supposed to direct. Glushkov was not an economist. He was a pure mathematician who, before the age of thirty, had solved one of the problems on David Hilbert’s famous list of unsolved questions, and had then turned his attention to computing machines. In 1956, he was put in charge of the laboratory in Kiev that had built MESM, the first working stored-program computer in continental Europe. By the early 1960s, that laboratory had grown into the Institute of Cybernetics of the Ukrainian Academy of Sciences, and Glushkov was its director. He looked at the Soviet planning system and saw a problem that was not political or economic in nature. He saw an information-processing problem, and he believed he knew what kind of machine could solve it. But he was not the first to say so, and the first man to say so had just been destroyed for it.

Anatoly Kitov was a colonel in the Soviet armed forces and one of the first people in the country to understand what a digital computer could do. He had run the Ministry of Defense’s first computing center and had written the first Soviet textbook on electronic digital machines. In the autumn of 1959, he did what no career officer was supposed to do: he wrote directly to Nikita Khrushchev, bypassing his entire chain of command. The letter later became known among Soviet computer scientists as the “Red Book,” after the color of its binding, and it stands as one of the earliest written proposals anywhere for a national computer network. In it, Kitov proposed that the military’s computing centers, which sat idle for much of the day, be connected into a single national network and used to run the civilian economy. During working hours, the machines would handle defense tasks; at night and during off hours, they would process economic data for Gosplan. One network, two purposes, and the entire planning apparatus of the Soviet Union would be automated in the process. He argued that the Soviet Union was falling behind the United States in computing and that a dual-use network was the only way to catch up without doubling the cost.

The letter reached the Central Committee. Then it was forwarded to the Ministry of Defense for comment. The generals read a document in which a colonel had gone over their heads to the head of state and had, in the process, written that the military’s management of computer technology was inadequate. They did not read a proposal. They read an accusation. A commission was convened to examine Kitov’s conduct rather than his idea. In 1960, he was expelled from the Communist Party, removed from his post, and pushed out of the military computing establishment he had helped build. Here is the first thing worth pausing on: the idea was not rejected because it was wrong. It was rejected because it threatened the people who would have had to approve it. That pattern is worth keeping in mind, because it would repeat itself at a much higher level for the next eleven years.

Glushkov knew Kitov, and he knew what had happened to him. He also knew the idea was sound. In 1962, he was summoned to Moscow to meet Alexei Kosygin, who was then first deputy chairman of the Council of Ministers and would within two years become the Soviet prime minister. Kosygin was an industrial manager by background, a former textile factory director who had risen through the economic ministries, and he understood better than most of the leadership how badly the planning system was failing. He asked Glushkov what it would take to fix it. Glushkov’s answer was a machine—not a single computer, but one connected system of computers covering the entire country. He called it the Unified State Network of Computing Centers, and the design would later be formalized under the name OGAS, the nationwide automated system for the collection and processing of information for the accounting, planning, and governance of the national economy. The name was bureaucratic. The design was anything but.

The network would have three levels. At the top, a central computing complex in Moscow. Below it, between one hundred and two hundred regional computing centers in major cities. Below those, as many as twenty thousand local computing centers located at factories, warehouses, collective farms, rail yards, and ports. Every level would be connected to the levels above and below it, and the whole structure would function as a single organism. The economic case Glushkov made was striking. He calculated that the initial investment in the network would pay for itself within a few years through reduced waste, better allocation of resources, and the elimination of the vast clerical workforce that existed solely to move paper between offices. But the most radical element of OGAS was not its scale. It was its architecture. Glushkov proposed that the network operate with a degree of decentralization, in which local computing centers would retain authority over local decisions and would communicate upward not raw data for central approval but processed information. The system was designed to support a transition away from rigid command planning toward something closer to a managed market, in which prices and production targets could be adjusted continuously based on real data rather than annual directives drafted in Moscow. In Glushkov’s vision, money itself could eventually be replaced by electronic accounting—a concept he discussed openly in internal documents and that later led some Western analysts to describe OGAS as an early blueprint for a digital economy.

Kosygin was impressed. By 1964, the project had secured backing at the highest levels of government, and Glushkov was given the resources to begin designing the system in earnest. The Institute of Cybernetics in Kiev expanded rapidly, attracting some of the best mathematical and engineering talent in the country. Work began on the central computing complex, on the regional center specifications, and on the software architecture that would tie the whole network together. The project was classified, and its details were compartmentalized, but its existence was known to a select group of planners, scientists, and military officials. American intelligence, which monitored Soviet computing developments closely, assessed the program as potentially transformative. A 1960s CIA report described the Soviet effort to automate economic planning as a development that, if successful, could alter the strategic balance of the Cold War by making the Soviet economy more efficient and more resilient. That assessment was, in retrospect, both a compliment to Glushkov’s ambition and a warning about the political obstacles the project faced.

The obstacles were considerable. OGAS threatened a vast array of established interests. The first was the military, which had never forgiven Kitov for his proposal and which viewed any civilian use of computing resources as a diversion from defense priorities. The generals controlled the country’s most advanced computing facilities, and they were reluctant to share them with planners. The second was the State Planning Committee itself. Gosplan’s authority rested on its role as the central clearinghouse for economic information. If a computer network could process that information more efficiently, Gosplan’s bureaucracy would shrink, and the officials who ran it would lose power. The third was the Ministry of Finance, which saw Glushkov’s ideas about electronic money as a threat to its control over the monetary system. The fourth was the network of regional party secretaries and industrial managers who benefited from the existing system’s opacity. A transparent, real-time economic network would expose inefficiency, falsified reporting, and corruption. For these officials, the current system was not a problem to be solved. It was a system they knew how to operate.

Glushkov understood this. He spent years building alliances, cultivating patrons, and framing OGAS as a tool that would strengthen rather than undermine the existing order. He argued that the network would give planners better information, not less authority. He promised that local managers would retain control over their own operations. He emphasized that the system would be built incrementally, starting with pilot projects and expanding only as experience justified. He gave lectures, wrote articles, and lobbied officials tirelessly. By the late 1960s, he had secured a powerful ally in Kosygin, who saw OGAS as a means of reforming the economy without resorting to the market mechanisms that were ideologically unacceptable. The project received funding, facilities, and personnel. It appeared, to those involved, to be on the verge of becoming a national priority.

But the political landscape was shifting. Kosygin’s influence within the leadership was declining as Leonid Brezhnev consolidated power. Brezhnev was a product of the party apparatus, and he viewed economic reform with suspicion. The conservative coalition that surrounded him—party ideologues, military leaders, and industrial managers—was deeply skeptical of any project that promised to redistribute information and authority. Glushkov’s network, for all its technical sophistication, required a level of institutional cooperation that the Soviet system was not designed to provide. And the project’s cost was enormous. Building a nationwide network of computers, communications links, and software systems would require resources that the Soviet economy could barely spare. By 1970, the estimated cost of OGAS had grown to a figure that alarmed even its supporters.

The decisive meeting took place on 1 October 1970. The Politburo, or a committee acting on its behalf, convened to review the project. Brezhnev was absent in Baku, and Kosygin was in Cairo. The men in the room were largely conservative, skeptical of Glushkov’s vision, and receptive to arguments that the project was too expensive, too risky, and too disruptive. Glushkov presented his case. He described the network’s architecture, its potential to reduce waste, and its role in modernizing the Soviet economy. He was, by all accounts, articulate and persuasive. But the discussion turned, as such discussions often do, to a practical example. One of the officials present asked about the chicken farms. The Soviet Union had recently launched a major initiative to increase poultry production, and the official wanted to know how OGAS would help. Glushkov explained that the network would collect data from farms, track feed and production, and help planners allocate resources more efficiently. The official listened, nodded, and then observed that the chicken farms were already being managed by a separate ministry with its own reporting system. If OGAS took over that function, the ministry would lose control. And if the ministry lost control, it would resist the project. The room understood the implication. The chicken farms were a microcosm of the entire economy. Every sector had its own bureaucracy, its own interests, and its own reasons to oppose a centralized network. OGAS was not just a technical challenge. It was a political one, and the political obstacles were insurmountable.

The project was not formally cancelled that afternoon. But the decision was made. Funding was frozen, then reduced, then redirected to other priorities. The central computing complex was never completed. The regional centers were never built at scale. The software architecture remained a blueprint. Glushkov continued to advocate for OGAS for the rest of his life, but the political will had evaporated. He died in 1982, and with him went the most ambitious attempt to computerize a national economy in the twentieth century. The Soviet Union continued to plan with pencils, paper, and abacus-like methods until its collapse in 1991. The inefficiencies that Glushkov had predicted—the growing burden of planning calculations, the inability to process information at the speed required by a modern economy—contributed to the stagnation and eventual failure of the Soviet system.

The story of OGAS is often told as a tragedy of technological promise thwarted by bureaucratic resistance. There is truth in that. Glushkov was a visionary, and his network was decades ahead of its time. But the story is also a case study in the limits of technology as a solution to political problems. OGAS was designed to solve an information-processing problem, but the problem it actually faced was one of power and incentives. The Soviet system was not designed to share information transparently, because transparency threatened the authority of those who controlled it. Glushkov’s network would have required officials at every level to surrender a degree of control, and they were not willing to do so. The chicken farms were not the reason OGAS failed. They were a symbol of the reason. Every sector of the economy had its own chickens, its own ministry, and its own reasons to resist. The machine could have processed the data. It could not have resolved the politics. And in the end, the politics won. The Soviet Union chose to keep its pencils and its paper, and the electronic brain that might have helped run the country was left to gather dust in a filing cabinet in Kiev.

Related Posts

Ancient DNA Sheds Light on Celtic Druids’ Origins

The year 8060 is a typo in the source material, but the story it introduces is one of the most enduring puzzles in European history. The scene…

The Difficult Life of the Forgotten Medici Girl

The story of the Medici family is one of ambition, artistry, and political cunning, a saga that helped shape the Renaissance and left an indelible mark on…

Archduke Franz Salvator of Austria | The Former Son-in-Law of the Habsburg Empire

The Habsburg Empire in its final decades was a place of extraordinary contradiction, a civilization at its cultural peak existing within a political structure that seemed increasingly…

Germany Had Armored Divisions in France — So Why Couldn’t They Stop D-Day?

Germany had roughly a dozen Panzer divisions waiting in the West in the spring of 1944, yet on June 6, only one of them drove toward the…

‘Y’all Nazi Gals Hungry’ | Texas Cowboys Gave German POWs Real Fried Chicken They Asked for More

The transport truck rattled to a stop outside the wooden gates of Camp Swift, Texas, on June 12th, 1945, kicking up a cloud of red dust that…

The Royal Jewels That Disappeared – Stolen, Recut or Lost for Centuries

The royal vaults of history were built to protect the most precious treasures of civilization, yet some jewels vanished anyway — through theft, revolution, betrayal, and simple…

Leave a Reply

Your email address will not be published. Required fields are marked *