This blue line on the graph shows gross investment by private domestic firms, while the red line shows net investment by private firms, both divided by GDP. You can see that from the 1970s and up into the 1990s, high levels of gross investment exceeded 14% of GDP. But since 2000, high levels of gross investment don't reach 14% of GDP. Interestingly, the drop-off in investment seems more visible in the red line showing net investment.
In the next figure, net domestic investment by private domestic firms is divided by gross investment: in effect, this calculation shows what percentage of total investment is actually adding to the capital stock, rather than just replacing earlier investments that have depreciated. The striking pattern is that from the 1960s up to the early 1980s, it was common for about 40% or more of total investment to be "net" or new investment. But since about 2000, it's been common for about 20% of total investment to be "net" or new investment, while the other 80% is replacing older capital stock.
The decline in net investment also shows up in government infrastructure investment, especially in the years since the Great Recession. Here's a figure from "If You Build It: A Guide to the Economicsof Infrastructure Investment," a useful overview of issues related to infrastructure spending by Diane Whitmore Schanzenbach, Ryan Nunn, and Greg Nantz (Hamilton Project, February 2017).
I haven't done a deep enough dive into the underlying methodologies here to see why the net/gross ratio for private investment is falling so sharply, or if some of these reasons may help to to explain the fall in net infrastructure investment. I have seen some discussion that part of the reason is that capital investment is more likely to be in the form of computers and software, which become outdated more quickly (given technological progress in this area) than, say, large machine purchased for industrial production in old-style plants. But there are other possible explanations. (If someone out there has dug down into the growing gap between gross and net investment and how it manifests itself in the Bureau of Economic Analysis statistics, please send me the paper or a link. I'd be happy to learn more.)
The decline in investment is bothersome in a number of ways. Investment in physical capital is one of the factors that over time raises productivity and wages. It's a little troublesome that 80% of gross investment is going to replace old capital, rather than add to the capital stock. And low investment is at the root of concerns about the possibility of "secular stagnation," which is a worry that the economy is headed for a slow-growth future because investment spending is likely to remain low.