The US National Debt Just Crossed $40 Trillion

It happened. On this site's model, United States government debt passed $40 trillion on August 17, 2026 — a threshold that did not exist in serious conversation a generation ago and that the counter on our homepage rolled through in the middle of a quiet Monday. As of August 18, 2026, the US figure stands at approximately $40.00 trillion, and the world total at approximately $117.03 trillion. Both numbers are computed from the same dataset that drives the live tracker, so every figure below can be checked against the raw file.

Our July mid-year recap flagged this crossing a month in advance, and put it at "around August 17." That is where the arithmetic landed. Here is what the number actually contains.

What crossed since the July recap

In the 31 days between our July 18 checkpoint and today, the model added roughly $161.2 billion to the US total and roughly $415.2 billion to the world total. Three lines were crossed in that window:

• World debt passed $117 trillion — around August 16, 2026. The sixth trillion-dollar line of the year, arriving on the same 41-day cadence as $115 trillion in late May and $116 trillion in early July.
• China passed $19.7 trillion — around August 12, 2026, on the narrow general-government basis the dataset uses. The true burden is considerably higher once off-balance-sheet local borrowing is counted; see our explainer on China's LGFVs.
• US debt passed $40 trillion — around August 17, 2026, roughly 138 days after the April 1 baseline of $39.28 trillion.

The $40 trillion number, unpacked

The US total grows in the model at $60,200 per second. That works out to about $216.7 million per hour, $5.20 billion per day, and roughly $1.90 trillion per year — which means the next trillion, from $40 to $41, takes about 192 days at the current rate.

Divided across a US population of 336.05 million, $40 trillion is about $119,000 of federal debt per resident, climbing by roughly $15.48 per person per day. Per resident is the cleanest of the per-person framings; per taxpayer and per household produce far larger numbers on far shakier assumptions, and we walk through why in our piece on per-taxpayer arithmetic.

Against the global picture, the United States now accounts for about 34.2% of all government debt on Earth in this dataset, though a smaller share of the growth — $60,200 of $279,686 per second, or 21.5% — because faster-growing borrowers elsewhere now add more each second than their stock of debt would suggest. Add China at about $21.30 trillion, the EU aggregate at about $16.09 trillion, and Japan at about $9.02 trillion, and those four blocs alone come to roughly $86.45 trillion, or 73.8% of the world total. The remaining 130-plus countries in the dataset share what is left.

An important caveat about the crossing date

This is the part most likely to be misread, so we will be blunt about it. The August 17 crossing is a modeled estimate, not an official Treasury announcement. No agency published a statement that day; nobody rang a bell. Our figures are educational estimates anchored to IMF World Economic Outlook baselines and extrapolated forward at a constant per-second rate, reviewed quarterly — they are not real-time reads of any government accounting system, and they are not certified government totals.

Real sovereign debt does not accrue smoothly. It moves in discrete steps as the Treasury settles auctions, redeems maturing securities, and adjusts intragovernmental holdings, and the official total can jump or fall by tens of billions in a single day. Different official measures also disagree with one another by design — gross versus net, general government versus federal, market versus face value — which is why our figure and any given headline figure will rarely match to the dollar. Our explainer on how the IMF measures debt covers those definitional gaps, and the full extrapolation method is documented on our methodology page.

So the honest statement is this: on a widely used international measure, extrapolated transparently, US government debt passed the $40 trillion mark in mid-August 2026. The exact hour is an artifact of the model. The magnitude is not.

Is $40 trillion the right thing to be alarmed about?

Not on its own. A nominal total tells you nothing without a denominator, and the economy generating the tax base grows in nominal terms too. The ratio economists actually watch is debt-to-GDP, which the dataset puts at 124% for the United States against a world figure of about 94%. That is high by American historical standards and unremarkable next to Japan's 207% — and Japan, as we have written before, is not a country markets treat as a default risk. Twenty-one of the sovereigns in this dataset carry debt above 100% of GDP.

There is also no threshold at which something breaks. The famous 90% line came from a study undone by a spreadsheet error, and the real determinants — who holds the debt, in what currency, at what maturity, at what interest rate — are country-specific. Our piece on why 90% isn't a magic line makes that case in full. The more useful thing to watch than the trillion-count is the interest bill, which is the fastest-growing line in the budget.

What is projected next

All of the following are projections from the current baseline and rate, not scheduled events. If borrowing accelerates or slows, they move.

• World debt is on pace for $118 trillion around September 27, 2026, at $279,686 per second and about $24.2 billion per day.
• US debt is projected to reach $40.5 trillion around late November 2026.
• World debt is on pace for $120 trillion around December 18, 2026.
• US debt is projected to reach $41 trillion around late February 2027 — roughly 192 days after the $40 trillion crossing.

Before any of those, the dataset hits its October 1, 2026 baseline review, when we reconcile the extrapolation against the IMF's autumn figures. That review is where reality gets to disagree with the model. If US borrowing ran hotter or cooler than the April baseline implied, the base figure and the per-second rate get restated, and every projection above gets revised with them — including, potentially, a retroactive shift in when the $40 trillion line was really crossed. We will publish the reconciliation either way, and we will say plainly if the model was off.