Foreign debt doesn't have to increase by the U.S. deficit if US currency used in foreign trade makes up the difference. The U.S. approach since 2009 has generally been to monetize the debt [1], having the Federal Reserve buy it and issue new money in its place. Normally this would result in inflation per the money equation PQ = MV; if you hold velocity V roughly constant, an increase in the money supply M either needs to be backed by an increase in real goods transacted Q, or it results in a higher price level P. Per the FRED graph, M2 money supply has been growing significantly faster than real GDP, and so we should've been seeing significant inflation over this time period.

But the dollar is used for 60-80% of international trade [2]. Many dollars don't stay in the U.S. to be transacted among the locals; they go abroad, and are held as foreign currency reserves or transacted between other nations for international trade. Under these conditions, the relevant Q in the money supply equation is the value of international trade denominated in dollars, which has been growing close to the increase in money supply. The dollars are getting soaked up for international transactions, preventing severe inflation at home.

I'm trying to model out what happens if the volume of dollar-denominated international trade declines, which seems to be the situation we're entering now and is the stated policy of the Trump administration. Certainly one consequence is that the stock market would crash by ~70-80%, which is the component of major S&P 500 companies earnings that are earned abroad. I suspect that this would trigger major chaotic effects (like WW3 or a revolution in the U.S.), that make future economic predictions irrelevant.

[1] https://fred.stlouisfed.org/graph/fredgraph.png?g=1UTPn&heig...

[2] https://www.brookings.edu/articles/the-changing-role-of-the-...