How a defensive alliance built an expansion architecture, and what George Kennan said would happen next.

January 9, 1990. Moscow. James Baker sat across from Mikhail Gorbachev in the Kremlin. Germany was about to reunify. Gorbachev needed to know what NATO would do with the new territory. Baker made an assurance. Not one inch eastward, he said. Not one inch would NATO expand toward Russia’s border.

Baker was not alone. The declassified documents from the NSA Archive show the same assurance was made by nine other Western leaders in the months that followed: Helmut Kohl of Germany, Hans-Dietrich Genscher, British Foreign Secretary Douglas Hurd, French President Francois Mitterrand. The same message, delivered in different rooms, with different words, amounting to the same commitment.

The assurances were not written into a treaty. That distinction would matter.

NATO now has thirty-two members. Nine of them share a border with Russia. The expansion includes every country from the former Warsaw Pact that could be admitted, and several that were once republics of the Soviet Union itself. The alliance that promised not to move one inch eastward has moved its eastern border to within two hours of Saint Petersburg.

The Mandate and Its End

NATO was founded in 1949 with a specific and explicit purpose. The North Atlantic Treaty created a collective defense arrangement under Article 5: an attack on one member would be treated as an attack on all. The Soviet Union was the named threat. The Warsaw Pact was its military answer. Europe was divided. The architecture made sense within that division.

On February 25, 1991, the Warsaw Pact formally disbanded. On December 25, 1991, the Soviet Union dissolved. The flag over the Kremlin came down. Gorbachev signed the documents. Fifteen new states emerged from the wreckage. The entire strategic rationale for NATO, the specific military threat it was built to counter, ceased to exist.

The threat ended. The institution did not.

NATO’s sixteen member states met, considered their options, and continued. This was not automatic. It was a choice. The question of whether a Cold War military alliance had any function in a post-Cold War world was genuinely open in the early 1990s. Some argued it should evolve. Some argued it should wind down. The institution itself argued it should stay.

In 1991, Boris Yeltsin sent a letter to NATO headquarters in Brussels. Russia wished to begin the process of joining the alliance. The response was the Partnership for Peace, a consultative program that offered dialogue without membership. The country that had been the reason NATO existed was offered a waiting room. It was never invited inside.

By 1995, Yeltsin had changed his position. NATO expansion, he said, was a threat to Russian security. The institution that had declined to include Russia was now preparing to expand toward it. The line between those two decisions is short and straight.

The Assurance and the Expansion

In February 1997, George Kennan published an opinion piece in the New York Times. Kennan was eighty-three years old. He had designed America’s containment strategy in 1946. He was the most consequential American foreign policy thinker of the twentieth century. His piece was titled ‘A Fateful Error.’

‘Expanding NATO would be the most fateful error of American policy in the entire post-cold-war era,’ he wrote. ‘Such a decision may be expected to inflame the nationalistic, anti-Western and militaristic tendencies in Russian opinion, to have an adverse effect on the development of Russian democracy, to restore the atmosphere of the cold war to East-West relations, and to impel Russian foreign policy in directions decidedly not to our liking.’

That was 1997. NATO expanded in 1999.

Poland, Hungary, and the Czech Republic became members in March 1999. Kennan watched. In April of the same year, NATO bombed Serbia without a United Nations Security Council mandate. It was the first time the alliance had used military force against a sovereign state. There was no Article 5 trigger. No NATO member had been attacked. An independent international commission later described the operation as ‘illegal but legitimate.’ The legal framework NATO had operated within for fifty years had been set aside.

In 1997, the NATO-Russia Founding Act was signed in Paris. It was designed to compensate Russia for the expansion it had been told would not happen. The act created a NATO-Russia Council and stated that NATO had ‘no intention, no plan and no reason’ to deploy nuclear weapons or large combat forces on the territory of new members. It also stated that NATO and Russia ‘do not consider each other as adversaries.’ The ink was barely dry when Poland, Hungary, and the Czech Republic were admitted two years later.

In 2004 came the largest single expansion in NATO history: Bulgaria, Estonia, Latvia, Lithuania, Romania, Slovakia, and Slovenia. Seven countries in one wave, including three former Soviet republics on Russia’s northern and western borders. The Baltic states had been part of the Soviet Union itself. They were now NATO members under Article 5, meaning any conflict involving them would commit the entire alliance.

Each expansion was described as defensive. Each expansion moved the alliance’s eastern border closer to Moscow.

In 2008, at the Bucharest summit, NATO stated that Ukraine and Georgia ‘will become members.’ The statement was non-binding, with no timeline. But it was written into the summit declaration. France and Germany had opposed it. The United States pushed it through. Russia’s objections were documented and dismissed. Four months later, Russia went to war with Georgia.

Four Wars. Not One Under Article 5.

NATO has fought four significant military operations since the Cold War ended. Bosnia in the 1990s. Kosovo in 1999. Afghanistan from 2001 to 2021. Libya in 2011. None of them was an Article 5 operation. Not one involved an attack on a NATO member state. Every one of them expanded the operational scope of an alliance whose founding document describes it as defensive.

Kosovo has been examined in detail by international legal scholars. The Independent International Commission on Kosovo, established with UN support, concluded in 2000 that the NATO intervention was ‘illegal under international law’ because it lacked Security Council authorization, but ‘legitimate’ because it halted atrocities. The legal framework NATO had operated within for fifty years had been redefined. The alliance that could only act in self-defense had acted in anticipation of a threat it defined itself.

Afghanistan ran from October 2001 to August 2021. The invocation of Article 5 after September 11 was the first and only time it was used. NATO committed twenty years and an estimated two trillion dollars. The Taliban, the same force NATO had entered to defeat, controlled most of the country within seventy-two hours of the final withdrawal. The government NATO had spent twenty years building collapsed in eleven days.

Twenty years. Two trillion dollars. Seventy-two hours.

Libya in 2011 was different in character but consistent in pattern. The United Nations Security Council passed Resolution 1973, authorizing a no-fly zone to protect civilians. NATO took command of the operation. Within weeks it was conducting strikes against Libyan ground forces, providing close air support for rebel fighters, and targeting Muammar Gaddafi’s compound. Russia and China had abstained from the Security Council vote rather than veto it. Both said afterward that they had been deceived about the scope. Russia’s foreign minister Sergei Lavrov said NATO had ‘grossly violated’ the mandate. The no-fly zone had become regime change. The UN framework had been used once and exceeded immediately.

After each of these operations, NATO did not contract. It expanded. After Bosnia and Kosovo, ten new members over five years. After Afghanistan began, seven more. After Libya, three more in the following decade. The institution that failed to achieve its stated objectives in four consecutive military operations grew larger with each failure.

The mandate expired. The architecture did not.

The RAND Blueprint

In 2019, the RAND Corporation published a report titled ‘Overextending and Unbalancing Russia.’ It was funded by the United States Army. The report’s stated purpose was to identify strategies that would ‘stress’ Russia’s military and economic systems, forcing it to spend resources it could not afford and make decisions it would rather not make.

The report examined multiple options. Among the most significant was arms support to Ukraine. The executive summary used precise language: ‘Providing lethal aid to Ukraine could exploit Russia’s greatest point of external vulnerability.’ The report acknowledged that this option carried risks. It acknowledged that Russia might respond with military force. It assessed that risk as manageable.

The document was written in 2019. Russia invaded Ukraine in February 2022.

This is not a causal claim. The RAND report did not cause the invasion. But it describes the institutional logic operating during the years between 2019 and 2022. Weapons began flowing to Ukraine in significant volumes. NATO training missions expanded. Infrastructure was built. Ukraine’s military capability, by any measure, increased substantially in the three years before the invasion. The alliance that said it was not a party to any conflict had built a significant part of the architecture that preceded it.

The report exists. It is publicly available on the RAND Corporation’s website. It was funded by the United States Army. It identified Ukraine as Russia’s greatest external vulnerability. And it recommended exploiting that vulnerability as a strategic tool three years before the war began.

The Account

Raytheon Technologies began 2022 with a market capitalization of approximately 128 billion dollars. By the end of that year, it was 155 billion dollars. Lockheed Martin started 2022 at 98 billion dollars and ended it at 127 billion. Both companies are primary suppliers of the weapons systems flowing into Ukraine under NATO coordination. Raytheon makes the Stinger missile. Raytheon and Lockheed jointly produce the Javelin anti-tank system. In August 2022, the US Army awarded the Javelin Joint Venture a manufacturing contract worth 1.3 billion dollars. Raytheon’s backlog of signed but unfulfilled defense contracts grew from 63 billion dollars at the end of 2021 to 77 billion dollars by mid-2024.

NATO Secretary General Jens Stoltenberg described the conflict in Ukraine in 2022 as a ‘war of attrition.’ The phrase is precise. A war of attrition does not end quickly. It consumes material at high rates. It requires sustained production. It creates sustained demand. The five largest defense contractors by revenue in 2023 were all American: Lockheed Martin, Raytheon, Northrop Grumman, Boeing, and General Dynamics. All five have seen defense order backlogs grow substantially since 2022.

Stoltenberg proposed that NATO take a larger role in coordinating the supply of weapons, ammunition, and training to Ukraine. He hosted meetings of defense officials from forty nations at Ramstein Air Base in Germany. The coordination of arms transfers to a country outside the alliance, fighting a war the alliance describes itself as not party to, is one of the more precise illustrations of what NATO has become. It is not the institution described in the 1949 treaty. It is something that has grown beyond that description while retaining the name.

The institution that coordinates the war is not the institution that profits from it. But the institution determines the conditions under which profit is possible.

European governments have been told to raise defense spending to two percent of GDP. For Germany, that required amending the constitution, a document written specifically to limit German military capability after 1945. The constitutional amendment passed in March 2025. The Bundeswehr, which had 495,000 soldiers at the end of the Cold War and fell to 183,000 by 2023, is rebuilding. The Netherlands eliminated its tank fleet entirely and is now reconstituting it. Each of these decisions represents a transfer: of public funds, of political priority, of European manufacturing capacity, toward a military posture designed within NATO’s framework and resourced primarily through American suppliers.

Jens Stoltenberg served as NATO Secretary General from 2014 to 2024. He joined the Bilderberg Group’s Steering Committee after leaving the position. The institutional network connecting NATO’s leadership to the financial and policy structures that benefit from sustained European military spending is not hidden. It is simply not discussed in the context of strategic decisions about whether to negotiate or continue.

What George Kennan Understood

George Kennan died in March 2005. He was one hundred and one years old. He had watched NATO expand three times since his 1997 warning. Poland, Hungary, and the Czech Republic in 1999. Seven more in 2004. Bulgaria, Romania, and others. He had watched Kosovo. He had watched Afghanistan begin. He did not live to see Libya, or Ukraine.

His warning was specific. He did not say that Russia would respond militarily out of irrational nationalism. He said that the expansion would ‘impel Russian foreign policy in directions decidedly not to our liking.’ He predicted the response would be a function of the action. He was describing cause and effect, not Russian character.

The counter-argument is that Russia’s actions in Ukraine are Russian choices, made by Russian actors, for Russian reasons. That is also true. Both things can be true simultaneously. An institution can create conditions that make a specific outcome more likely, and an actor within those conditions can still choose how to respond. Kennan’s point was not that expansion would make war inevitable. His point was that expansion would make hostility rational, from Russia’s perspective, in a way that it had not previously been.

Russia did not become a threat because of what it is. Russia became a threat because of what NATO did. Kennan wrote that in 1997. No one who mattered was listening.

The Baker assurances were not written into a treaty. That has been NATO’s legal defense ever since. Verbal assurances are not binding under international law. This is technically accurate. It is also a description of exactly the kind of behavior that makes agreements between states difficult to sustain. Russia was told one thing. It watched something different happen. Sixteen times.

In March 2022, six weeks into the Russian invasion of Ukraine, negotiators from both sides met in Istanbul. By the end of the month, they had produced a framework document. Ukraine offered to remain neutral and not join NATO. Russia offered to pull back to the lines existing before February 24. Both sides described the talks as substantive. In April, British Prime Minister Boris Johnson visited Kyiv. The talks ended shortly afterward. Arakhamia, the lead Ukrainian negotiator, said Johnson had told the Ukrainian side not to sign anything and ‘let’s just keep fighting.’ Johnson denied it. The talks did not resume.

Whether Johnson’s visit was decisive or incidental, the outcome is documented: a negotiating framework existed, and the war continued. The institution whose expansion had contributed to creating the conditions for the war had no mechanism for ending it. It had a mechanism for sustaining it.

The Kennan analysis was shared widely in 1997. It was not unique. Jack Matlock, the American ambassador to the Soviet Union during its final years, opposed the expansion. William Burns, who served as ambassador to Russia and later became CIA director, wrote in a 2008 cable released by WikiLeaks that NATO membership for Ukraine was ‘the brightest of all red lines for the Russian elite.’ He wrote it as a warning, not as an endorsement. The warning was forwarded up the chain. The Bucharest summit statement was issued four months later.

NATO was built to deter an attack that never came, against a military bloc that dissolved itself, in defense of a continent that financed the deterrence for forty years without being invaded. That is one account of the Cold War.

A second account begins in January 1990, in Moscow, with a promise that was not kept. It continues through thirty-two memberships, four wars, one RAND report commissioned by the U.S. Army, and a negotiation in Istanbul that was more promising than anyone in Brussels wanted to acknowledge.

George Kennan said in 1997 that the expansion would be a fateful error. He was specific about the mechanism: it would ‘impel Russian foreign policy in directions decidedly not to our liking.’ He was not predicting Russian aggression. He was describing the architecture that would make aggression logical.

The architecture produced the outcome. The institution that built the architecture is still expanding.

Sweden became the thirty-second member in March 2024. Sweden shares no border with Russia. Sweden had maintained neutrality for over two hundred years. The decision to join was made in the context of the Ukrainian war. The war that began, in part, because of the conditions the alliance had spent thirty years building.

Finland joined in April 2023. Finland shares a 1,340-kilometer border with Russia. The two countries had maintained a working relationship since 1948 precisely because Finland understood the geography. That relationship, built over seventy-five years of careful coexistence with a nuclear neighbor, was ended in fourteen months. It is now an Article 5 commitment. Any conflict on that border triggers the full mechanism of the alliance. The institution that began with sixteen members in 1949 now covers a border with Russia that extends from the Arctic Ocean to the Black Sea, interrupted only by Belarus.

The bill for that architecture is now distributed across thirty-two defense budgets, a German constitutional amendment, a Dutch tank fleet being rebuilt from scratch, and the families of the approximately five hundred thousand people estimated to have been killed or wounded in Ukraine by the end of 2024. None of them appear in the RAND report as a line item.

The institution answered the question in 1990. It said it would not move one inch eastward. Europe is still paying for what happened next.

Jerry writes The Manifest Archive, a forensic analysis of the institutional structures that shape geopolitics, history, and power. Published on Substack and Medium.

Related from The Manifest Archive