Good grief! ARPANET was indeed a huge project at the time. All the primary components of the internet came from very large government projects. Those who argue otherwise are mostly ignorant of government procurement practices (almost all governments, in fact). They regularly establish overall projects and fund tiny pieces one at a time.
So, depending on whether you view the overall project or only a given subcontract gives you wildly different views:
So, IP, mouse, domain names, telecomm form T1 to Giga-speeds, etc, every single one was a tiny piece attached to a huge project.
As for funding after initial development you need only look at geophysics, oil depletion allowances, military security for oilfields, development and buildout of the internet, the GPS system (Glonass too, actually). Those have gigantic government subsidies to promote private sector profits. Then consider carried interest, and MACRS.
One of the monumental logical failings of those who oppose BEV adoption subsidies and government support of buildout for charging infrastructure totally ignore all the subsidies devoted to other subjects.
I suggest anybody who disagrees with the foregoing might do some modest investigation of how much government subsidy is devoted to building and supporting sports stadiums in the US. Once someone does that they'll find out that BEV direct subsidies and free public charging country wide all together will actually be LESS than the commercial sports subsidies.
I am not necessarily arguing against the commercial sports subsidies. I do argue that BEV support is far more of a benefit to the populace than are the sports, which could be self-funded anyway.