Or you know - have the government own the poles and sell access to anyone who wants to run lines. Private roads are impractical except in limited circumstances (that require a huge number of public roads to enable). Infrastructure like this is similar.
Under federal law, utility pole owners are required to sell access to poles on a non-discriminatory basis. There is no doubt that Google is allowed to run their fiber on the poles. The lawsuit is over provisions of the federal law that provide for a notice/"make ready" procedure for new attachments. Under that procedure, Google would file a request to use the pole, and the pole owner would have 145 days to notify everyone else using the pole and move equipment as necessary in order to get it ready for Google's attachment. Here, the municipality passed on ordinance allowing Google (and those similarly situated), to bypass that procedure and move other companies' equipment already on the poles in the process of putting up its own.
That can work but you have to do a much better job of outlawing all forms of government-protected monopolization and its half-way equivalents. Otherwise what you end up with, are local governments that want to cut deals for extra cash with AT&T et al. Local governments are supposed to not be allowed to cut monopoly deals today, but they still conspire with big telco to rig the game any way they can.
People always like to say private roads are impractical, but that is 1970's thinking. Today all we need is a common standard for toll readers like most states have for their internal highway toll system, and people could drive on toll roads and be automatically billed, and your GPS can factor in the toll costs in your travels (just as they do now - in most GPS navigation apps, you can elect to avoid toll roads).
The comparison to utility lines is not particularly relevant, though. If I want to travel over a road, I just need to bring my car and do it once. If I want to run gigabit fiber over phone poles, it is building a road on a road - the fiber cable is itself a road, build on the infrastructure transit center of phone poles.
The more apt comparison is that I want to send data over said fiber channel - it is the functional equivalent to wanting to drive my car along a road, the only real difference is the marginal cost of that traversal is so insignificant on the fiber line (the cost of beaming light along it in a fixed data path) that you may as well disregard it, whereas cars have significant per-distance and per-usage costs associated.
But roads and coax / twisted pair / fiber lines are really great comparisons to roads. They have usage limits, they can become congested, but when they are underutilized they are effectively wasted. It seems like something private enterprise would actually be good at, if not for the fact possessing the road or cable line grants yourself an untenable market position giving yourself a monopoly. But public transit authorities demonstrate very poor ability to adapt to consumer demand for expanded roads where there is great congestion, because they aren't profiting off road traffic and are not incentivized directly to fine the maximum tradeoff of value against consumer satisfaction - there are many degrees of separation between the funding of road construction, the maintenance of the road system, and the consequences of your "customers" being dissatisfied. Usually, the only ways dissatisfied customers who dislike the public road network can demonstrate their demand is by moving or voting out politicians - if those were the same barriers to faster Internet speeds, we would probably be as bad off as we are now with telecom monopolies.
You buy the land from individuals and pave it. Eminent domain is relatively rarely used in most road construction - at least in the Northeast, almost every road here used to be a dirt trail two hundred years ago.
There is a natural equilibrium between consumer resistance to selling their land and market demand for roads. If you did not have eminent domain being used to compel home sales for new roads, you get this chain of events:
1. Existent road usages increases and grows congested.
2. Private road companies start offering residents to buy their lots to build new roads, because there is now a profit motive off building new roads, and one of two things happens:
A. They can build the road, increasing capacity, making money, and spurring economic growth.
B. They cannot build the road, and the area withers as businesses and people migrate away to avoid the congestion.
Either way, congestion solves itself, and for the landowners the optimal time to sell would be to sell to a road company - it means your land is at peak value, when demand for new infrastructure is high enough to justify investment, which means your property is as valuable as it is going to get.
It is important to remember when talking about things like private roads that we look at it through the biased lens of a society that has been erected on public roads, where you get told how much you get when your house gets eminent domained and the best you can do is appeal to a court that is also a part of the same state that took your land.
I would also imagine that a lot of homeowners associations would emerge naturally, that make it a stipulation of living in the neighborhood that in the event of sufficient demand for new road construction you be willing to sell at some multiple of inflation-adjusted property value, to guarantee that everyone has some price rather than having holdouts who just refuse for any amount. But it would still be your decision to move into such arrangements, and a lot of people probably would not, but businesses would likewise strongly consider the growth prospects of neighborhoods without strong homeowners association sale bindings on property in the area and avoid regions where they know it is unlikely infrastructure expansion could organically keep pace with demand.
