In 1940, the United States placed Japan in a difficult position. Japan was a country with very limited resources. It had to import oil, steel and other goods from other Asian countries and, to some extent, from the United States. To secure access to these resources, it had tried, years earlier, to build an empire. The U.S. used its economic power to block the sale of oil from what we now call Indonesia, to name just one country, and refused to sell steel to Japan. The U.S. feared that a Japanese Empire would threaten U.S. military command of the Pacific, making the homeland vulnerable to Japanese military power.
This culminated in a U.S. blockade of Japanese imports critical to its industrial survival. Japan could either capitulate to the U.S. or go to war. Tokyo chose war, attacking Pearl Harbor in 1941 in an attempt to cripple U.S. power in the Pacific and force the U.S. to negotiate a new understanding with Japan. Given the antiwar sentiment in the U.S., this was not an outlandish thought, but it was an incorrect one. It’s reported that the commander of the Japanese Combined Fleet, Adm. Isoroku Yamamoto, who had spent time in the United States, opposed the attack, believing it would not lead to negotiation, but was overridden.
That Japan opted to fight was, I think, broadly predictable through geopolitical modeling, as was the U.S. decision to wage total war. How they would fight the war, and thus how they would end it, was far less predictable. The outcome depended on what I call engineering the details of geopolitics – relative power based not only on resources but on how those resources are used and deployed. This was the realm of detailed intelligence gathering and its application by the military. The engineering is much harder to predict.
I offer this example to contextualize the war in Iran. That there would be a war was, to a degree, predictable, as is the desired outcome of each side. How the war plays out, and who will achieve their desired outcome, is much less predictable, as war is the domain of the engineers, and how they will choose to act is much more complex and, given the nature of choice, unpredictable.
I have argued that the threat of nuclear power in Iran was a fundamental threat to the U.S. based on the ideology of the Islamist groups that were operating in Iran and outside of Iran with Iranian support. The likelihood of a nuclear attack on the U.S. was always extremely low, but the stakes were extremely high. The U.S. view on this had been confirmed by the initial strike in June 2025 on Iran’s nuclear capability. Given its limited success, I expected another, broader attack on Iran. How that strike would occur was a matter for the engineers, intelligence organizations’ military strategists, soldiers with tactical training and, ultimately, political leadership.
What could not be predicted was what particular strategy would be chosen, and it now appears as though the strategy was based on intelligence about Iran’s military capability and internal political and economic situation. Central to it was the fact that there had been massive political unrest in Tehran. Large protests attended by those hostile to the government were interpreted as a sign that the government was weak and regime change would be welcome. This explains President Donald Trump’s call for Iranians to rise up against the regime.
The first phase of the war involved decapitation strikes that took out several Iranian leaders, but the Iranian public has not heeded the call. Meanwhile, Iran launched a flurry of drone and missile attacks on U.S. bases (and some national infrastructure) throughout the Middle East.
So either the intelligence community made a fundamental mistake or Trump disregarded the intelligence. I think the strategy illustrates a fundamental misunderstanding of the reality of how Iran operates and, particularly, the composition of the regime. The civilian leadership that was destroyed and replaced with another civilian leadership was the formal government of Iran, but it was not in operational control of Iran. Control largely rests in the hands of the Islamic Revolutionary Guard Corps. The destruction of the civilian regime did not paralyze Iran because the IRGC has a great deal of autonomy, is heavily involved in the economy, and exercises control over the actions it takes. As a result, the IRGC responded to the initial attacks quickly. In some very real ways, the regime governing Iran was the IRGC, and the destruction of the civilian regime in no way crippled or even weakened the government. The Japanese had assumed that the attack on Pearl Harbor would force the U.S. into negotiations. Instead, it forced the United States into a massive military response that ended at Hiroshima and Nagasaki.
The IRGC employed what’s known as a mosaic model of defense: Command structures are organized in such a way that if the central command were crippled or unable to communicate, regional commanders were empowered to take actions based on preplanned scenarios. Therefore, destroying the civilian government left the IRGC, or parts of it anyway, to continue the war as it saw fit.
In addition to its kinetic attacks, Iran also closed or at least limited access to the Strait of Hormuz, an indispensable thoroughfare for oil produced in the Gulf states. This has led to a massive global economic crisis, as oil prices rose dramatically, creating different rates and global shortages and, therefore, inflation.
This leads to something Trump had pledged to avoid. Clearing the strait presumes that boots on the ground will be necessary to keep enemy ground forces out of range, and that naval assets will need to clear the strait from naval interdiction. One problem is that Iran has the capability of stationing anti-shipping drones vast distances from the strait. They can operate with intelligence, if not satellite imagery, and then with firsthand observations from well-defended positions nearby.
To solve this problem, air intelligence and strike intelligence must be used in concert, and given the value of shipping and the cost of insurance, that data must be near-perfect to open the strait. This can be achieved only by inserting ground forces around the strait, with observers potentially much deeper, to secure a very large area from drones. On-the-ground observers are also in more danger and less efficient since they don’t see as broadly as satellites. Obviously, aircraft can be used to strike at drones as well, but it is possible to keep drones camouflaged and then launch them when needed. And because it would need to cover a large area and have near-perfect precision, a ground operation would require a very large force.
This, in turn, creates fundamental political problems for Trump, one of whose promises in the election was to avoid endless ground wars. It’s possible that none of these challenges will emerge, but they may, as could other unexpected challenges that trigger either an economic crisis, an extended war, or both. And it raises an old question: whether airpower on its own can defeat a widely spread enemy fighting on familiar terrain.
There is one last variable in this equation: the Artesh, Iran’s conventional military force that boasts more than 400,000 troops. It is a secular force committed to the defense of Iran. It is not Islamist but still Iranian, and has been targeted by limited U.S. and Israeli attacks. Iran does not seem to have called it into action. Its future actions thus increase the level of uncertainty in the war.
Geopolitics could have predicted that the Japanese would go to war with the U.S., but it could not specify how the war would be engineered, nor could it forecast its outcome. The same can be said for the Iran war. Geopolitical models could predict that there would be a war, but they cannot predict the strategies and tactics that will be used to fight the war, and therefore are not in general able to predict its outcome. That is the domain of military intelligence.
