How is this even a question? OF COURSE the ISPs are left to pay for "new internet investment", that is a material cost of doing business. In order to reach a subscriber, I have to run a cable or build a tower, which requires some capital expenditure.
For a mature/operational ISP, capital is usually obtained by charging the subscribers a monthly fee. The monthly fee covers things like infrastructure and the operating expenses, maintenance and upgrades thereto, labour costs for the people to operate, maintain and upgrade the network, support, service fees for bandwidth and/or usage and allows (usually) an amount usually referred to as "profit".
We know from their published numbers that most ISPs are making fairly obscene profits (90% is commonly quoted, but break it down and it's not always quite that high), and many of them are more than happy to take from the USF BUT users are not really seeing much benefit - to the point where some states are bringing lawsuits against telcos (WV vs Frontier springs to mind right now). Profit and availability of capital therefore is not really an issue in this industry so the question becomes one of what makes the ISPs believe that they are seemingly exempt from having this cost center in their business model? What makes them think they can charge twice for what is essentially the same transaction?
Sure, it would be silly to expect a leased line on dedicated infrastructure at the prices paid for typical residential service BUT to the same end bandwidth itself really isn't that expensive. Consumers will find ways to use it and this leads to a continual upgrading process, there is no question, but again, part of the cost of doing business as a middleman is buying sufficient product from your suppliers to be able to distribute to your clientele. You get more clients, you need to buy more product to account for that. In any other industry, they would get sued for breach of contract for not having enough (or the wrong) product.
Many will argue that "Hey, $ISP built this network so it's theirs to do with what they wish" but the reality is that many networks were taxpayer funded and so it goes without saying that if something is taxpayer funded, it absolutely should NOT be "private property" - they may have the contract to build/maintain/upgrade it, for which they get paid, but the same goes if I have a contract to build/maintain/upgrade something for which I get paid... just because I built/maintained/upgraded something and got paid for it, doesn't make it mine: it remains the property of the entity which paid for it.
Perhaps a better analogy is building a house. Unless you pay for it in cash, you probably have a mortgage. Yes, it is yours in title but the bank still owns it, even if you built it with your own two hands. And if you don't maintain it, the city council might have something to say about it. And if you don't periodically upgrade it, you may no longer be in compliance with one or more regulations. And if you stop paying for it, soon enough it'll no longer be yours.
The private property issue though does murk up the argument in the two paragraphs above, but there is some fairly fundamental differences, namely that the infrastructure in question is rarely ON what could really be considered private property, by which I mean, cables may go across private back yards, up private driveways and across private fields, but houses are not linked up in a serial, with each house relying on the previous house to remain connected. The ISP infrastructure in your house is for your household (which you paid for with your install fee, but that's another thing altogether), and has no place in the supply of services to your neighbours or neighbourhood and as such unless there is a problem with *YOUR* service, a technician doesn't require specific permission from any one individual to physically access the ISP equipment that supplies an area.
Given this, it would seem that some of these US ISPs want to have their cake and eat it too: public funding for building, maintaining and upgrading their (own, private, non-shared, closed) networks AND the ability to choose whether to build/maintain/upgrade their network and infrastructure WHILE charging consumers out the arse for all of the above, and profiting from it by offering services that may not be fit for purpose (leaving old infrastructure to decay while customers are still using it, charging for services not received, not upgrading congested peering links or finding other ways to reduce congestion).
Another popular argument is that of population density. This is a farce, of course - a city with 1 million people in the US is basically the same as a city with 1 million people on any other continent, so there is no reason a city in the US can't have European quality Internet access - yet I read daily of subscribers struggling to wrangle even a few megabits out of their connections.
Plus, the middle mile in the US is a very competitive market (much more than the last mile market) - as a city in Arizona recently found out, often we're talking about 1 or 2 cables to an area which is then shared by multiple providers. So even though the initial build cost isn't shared, it's not like any individual ISP is footing the entire bill over the entire life of the cable (even though they might claim they are) but whoever builds the cable has multiple customers buying large quantities of bandwidth or even wavelengths, all on the same cable, making such endeavours entirely viable and reasonably profitable.
Long story short: "new Internet investment" is part of the cost of doing business in this industry. Quit whining and deal with it. Frankly, I'm surprised that the government hasn't (or isn't getting ready to) repossess or put a lien on all the taxpayer funded networks so that you can get the types of systems we have overseas, i.e. shared/open/unbundled last mile access networks on which any ISP can offer access (which, frankly, I'd like to see happen in the US).