The US Army began using metric in World War I. Specifically 1918.
The American Expeditionary Force needed to make use of French intel, maps etc. Maps being the most critical since you needed to use kilometre-based grids for artillery fire.
The General Order of the United States Army on January 2, 1918 stated the AEF had adopted the metric system. This resulted in an immediate production of metric training literature for the army, the most prominent being:
So, there are a lot of interesting points and good questions being asked in this thread, but I feel like the overall consensus seems to be that the US Army switched to metric during World War One and never looked back. Thats not, precisely, accurate. While many organs of government over the years have pushed the metric system, it hasnt been accepted across America. So instead what you have is an uneven system, and this is also true of the military. Somethings are measured in Imperial measurements and others in metric, and why one gets measured one way and why another doesnt really doesnt make much sense.
Furthermore, because most Americans use the Imperial system, the units officers tend to use most frequently are those which they understand. Officers even today used pounds, miles, feet, and inches because its something that every American soldier understands naturally.
One great example of why the US Army has stuck to the metric system can be found in the MBT-70 tank development program. So, basically the MBT-70 was a tank that was developed jointly between the US and the Federal Republic of Germany. Teams in the US and in German co-developed technologies which went into the tank. But there was an obvious problem. After World War Two, German industry had been ravaged, especially in terms of its armaments industry. When they were rebuilding, the Germans had the opportunity to rebuild their machinery along the most modern lines, and specially it enabled a quick and easy shift towards the metric system. By the 1960s, most German industries had shifted to the metric standard, and it was a source of national pride. But in the US, the Imperial standard was the longstanding king.
Early on, German and American designers proved willing to draw up designs which were included both measurement systems. However, a factory designed to construct in one system could not easily switch over to the other. Factories in the United States, designed to produce items in inches, would need a completely new and expensive set of tools and machines to produce metric components. The same was true in Germany factories geared for the metric system. On March 9, 1964, at a meeting of the Program Management board the American Program Manager General Welborn Dolvin and German Manager Dr. Fritz Engelmann broached the subject of the threaded fasteners. The only thing they could agree on was the need to send the issue up to their bosses, their respective Secretaries of Defense.
For both Dolvin and Engelmann, the threaded fastener issue really represented a matter of national pride, rather than the goal of achieving a streamlined and logistically unified tank. The United States saw the imperial scale as the dominant system in western weapons production, weapons export, and civilian goods across the world. They argued that more goods were produced, worldwide, using the imperial system than the metric. Additionally, the American weapons which made up the bulk of the German Army all used imperial measurements. American industry too argued that the expense of switching from imperial to metric represented a significant cost that would not be incurred in Germany if they accepted the imperial standard for the MBT-70. For Germany, the problem was the exact reverse. In rebuilding after the destruction of World War Two, German industry had made the very difficult switch from imperial to metric. Even though the country and the military was used to, and to some extent dependent on, American supplies built on the imperial system, the von Hassel maintained that switching back to the imperial system would be a tacit admission of the inferiority of the metric system. For both nations, national pride was on the line in selecting a measurement system. Neither McNamara nor von Hassel could agree on which system to use for the project. A ministerial meeting held between May 9-11 acknowledged the issue of the threaded fasteners, but failed to make any serious headway, and the problem was delegated back to the Program Management Board, where it continued to divert time and energy away from the other problems of the MBT program. The issue was brought up at an interim meeting of the Management Board held in late June 1964. There, the Dolvin and Engelmann were again unable to solve the fastener issue, and again cleared it from their docket by referring the issue back to the ministerial level.
McNamara and von Hassel proved unable to remedy the fastener standardization issue. On June 26, three days after the conclusion of the interim meeting of the Program Management Board, the Managers jointly agreed to refer the issue back to McNamara and von Hassel. The Ministers met again to discuss the MBT program on November 13. They discussed a range of issues, including the location of the Joint Engineering Agency, the tank's military characteristics, and the threaded fastener issue. On the fastener issue McNamara and von Hassel were again unable to reach a satisfactory decision regarding which system to use. Instead, they concluded that they would completely table all discussion of the fasteners pending the result of a series of American and German studies focusing on which system would be best for the MBT project. The issue would not be raised again until May 1965. On May 19, 1965, over a year after the Program Managers first raised the threaded fastener issue, McNamara and von Hassel met to discuss the results of the German study concerning which fastener system to use. Following the results of the study, and given the pernicious and serious nature of the fastener issue, McNamara and von reached a compromise solution, which would preserve the pride of both countries. They agreed that each country could use its own system in components constructed domestically. Where components were joined together, or to the hull of the tank itself, they would use the metric system. At the time, this was considered a major concession to the Germans. However, the compromise solution solved none of the issues at play in the fastener debate. The goal of the joint MBT project was to create a tank which was logistically similar and simple. By allowing components built in both systems, industry contractors would still have to make the expensive investment in both metric and imperial tools. Repair crews had to use two sets of tools and two sets of parts to perform even simple repairs. The compromise solution satisfied no one, increased the complexity of the MBT-70, and solved none of the issues really at play. Unfortunately, time had run out to debate the issue, as the prototype stage was rapidly approaching, and finished plans were needed to advance the program.
So, as you can see there was a lot at stake in deciding which kind of system to use, metric or imperial. But because of the bastardized solution chosen by McNamara, it prevented the acceptance of one standard or another within the larger framework of the NATO alliance. In fact, because of the unsuccessful conclusion of the MBT project and the difficulties of cooperative development, I would argue that this event helped prevent the acceptance of the metric system during the closing years of the Cold War. Rather, the use of the metric system is largely confined to areas utilizing new technologies, or in a very select set of cases (like artillery sizes) which are based partially on tradition and partially on some actual practical benefit.
Sources:
Welborn G. Dolvin, Lessons Learned: Joint International Program Management For the US/FRG Main Battle Tank (Rock Island: U.S. Army Management Engineering Training Center, 1966)
"Main Battle Tank 70 Chronology," Project Management Planning Files, Office of the Project Manager for the Main Battle Tank, 1972; Box 5, RG 544, "Records of the Army Material Command," National Archives, College Park, Maryland
"13th PMB MTG 9-11 October, 1964"; Project Management Planning Files, Office of the Project Manager for the Main Battle Tank, 1972; Box 1, RG 544, "Records of the Army Material Command," National Archives, College Park, Maryland
Follow up:
Did this confusion (if any) contribute to the establishment of NATO's standardization system?