Here's a summary chart of where federal spending levels as a share of GDP would be in 2037 under different long-term budget plans drawn up by different think tanks.
Something you'll note is that there's substantial disagreement about the appropriate level of 2037 defense spending. Something else you'll note is that nobody at any of these think tanks has any idea whether the current form of government in the People's Republic of China will still be in place in 2037. Nor do they have any insight into how future rounds of nuclear arms control talks with Russia will go. They don't know what military technologies will be invented or what they'll cost. They don't know whether India will speed up to a China-like level of economic growth for a couple of decades or slow back down.
Not because they're dumb, but because forecasting is hard.
Nobody in 1969 knew the Cold War would be over in 1994. And nobody in the Budget Office in 1944 knew that two nuclear armed superpowers would be locked in a Cold War in 1969. Nuclear weapons didn't even exist. It would obviously be absurd for the country to be sitting around in 1969 having its military purchasing decisions driven by a budget forecast produced near the end of World War II just as it would be absurd for our 1994 defense spending to be driven by Vietnam-era thinking about military needs.
That's not a knock on any of the think tankers working on this stuff. Just an example of what kind of foolhardiness takes root in the deficit hawk community. When deficits are a problem—when excessive borrowing forces you into a sharp tradeoff between crowding out of private investment and inflation—they're a problem right away and you have to reduce borrowing right away. These airy speculations about the situation decades in the future are never the issue, because nobody can see the future and if they did they'd have better things to do with that knowledge than write think tank reports.