In this paper, we propose a new systemic risk indicator to measure the distance to the extreme losses of a financial system. Our indicator is based on cross-sectional concomitant VaR exceptions (Co-Exceptions) observed at a daily frequency, which are then converted into a weekly time series with only the maximum values to apply extreme value models. A set of 95 large U.S. financial institutions is used to run the empirical analysis over the last 20 years to check the real-time ability of our framework to predict significant financial crises, such as the Great Financial Crisis of 2008, the sovereign debt crisis of 2010 or the COVID lockdowns of 2020. Our systemic risk indicator identifies accurately this surge in systemic risk and provides additional information compared to the VIX indicator or the Value-at-Risk of the market. Finally, we show that this new measure of financial instability is explained by macroeconomic variables, such as industrial production and unemployment which have a positive impact, whereas the consumer price index, interest rate, and federal funds rate have a negative impact.