What is the original source for Vico's reference to Varro counting 30,000 pagan gods?

I have no information about where a figure of 30,000 gods of Latium might have come from. But here is some information that gives some perspective on the numbers of gods.

I may note that there are a great number of Hindu gods who have temples and worshipers, and other gods mentioned in various sources. The total number of Hindu gods is uncertain.

Thirty-three divinities are mentioned in other ancient texts, such as the Yajurveda,[114] however, there is no fixed "number of deities" in Hinduism any more than a standard representation of "deity".[115] There is, however, a popular perception stating that there are 33 crore (330 million) deities in Hinduism.[116] Most, by far, are goddesses, state Foulston and Abbott, suggesting "how important and popular goddesses are" in Hindu culture.[115] No one has a list of the 33 category goddesses and gods, but scholars state all deities are typically viewed in Hinduism as "emanations or manifestation of genderless principle called Brahman, representing the many facets of Ultimate Reality".[115][116][117]

This concept of Brahman is not the same as the monotheistic separate God found in Abrahamic religions, where God is considered, states Brodd, as "creator of the world, above and independent of human existence", while in Hinduism "God, the universe, human beings and all else is essentially one thing" and everything is connected oneness, the same god is in every human being as Atman, the eternal Self.[117][118]

So it is often said that there are millions of Hindu gods, which are often all considered aspects of one god or of the universe.

In Japanese Shinto there are many Kami, more or less gods or spirits.

There are considered to be three main variations of kami: Amatsukami (天津神, the heavenly deities), Kunitsukami (国津神, the gods of the earthly realm), and ya-o-yorozu no kami (八百万の神, countless kami). ("八百万" literally means eight million, but idiomatically it expresses "uncountably many" and "all around"—like many East Asian cultures, the Japanese often use the number 8, representing the cardinal and ordinal directions, to symbolize ubiquity.) These classifications of kami are not considered strictly divided, due to the fluid and shifting nature of kami, but are instead held as guidelines for grouping them.[3]:56

So one category of Japanese gods and god-like spirits is described as literally eight million, but figuratively as a vast, uncountable number.

So, since ancient Roman polytheistic religions were vaguely similar in some ways to Hinduism and Shinto, I find it easy to believe that some polytheists in the Roman Empire may have believed there were tens of gods, and others might have thought that there were hundreds of gods, and others might have thought that there were thousands of gods. And possibly some polytheists in the Roman Empire may have estimated that there were millions of gods.

I think perhaps this comes from changing definitions. Today, we tend to see two sharply-differentiated categories: natural and supernatural. We tend to think of gods as extremely powerful beings -- even most polytheisms today have a school of thought that argues that the individual gods are all facets of a single godhead -- and we have a category of minor supernatural beings which aren't gods: angels, vampires, elves, etc., etc., etc. (We'd never call a vampire a god, for instance.) This organized approach to the supernatural grew alongside our growing understanding of nature which, after all, started with cataloging and characterizing all of nature.

The ancient Roman religion was not like that. To start with, its gods -- spirits, whatever -- were not personalized. They weren't people, they we're not individuated beings, but rather were vaguer, ill-defined spirits of this or that -- Rome itself, War, Healing, Portals, etc., all the way down to spirits which were thought to be associated with every crossroads and every home.

One writer described the early Roman religion as a barely organized morass of superstitions.

Contact with Greek philosophy and the more advanced Greek civilization forced the Romans to think about their religion and to begin to systematize it. The Greeks had already started identifying their gods with various foreign gods, so they decided that the Egyptian Ammon and the Etruscan Tinia being the same as their Zeus. And in time, Roman Jupiter was added to the equivalences.

The Romans wound up with a wonderful mischmasch including equivalents of most of the Greek gods and also all with their ancestral spirits of each home and each crossroads and each just-about-anything-else.

(Note also that the Roman practice of deifying Emperors makes sense only with a fairly broad definition of a god.)

By Roman standards, were all these spirits gods? I don't think that's a meaningful question. They were less powerful than Jupiter, to be sure, but -- in early- and mid-antiquity, anyway -- there was no clear distinction. (Late antiquity is quite different, with the neo-Platonists attempting a systematic theology of the Roman paganism in hopes of offering an alternative to Christianity. I know little about that phase.)

So how many gods did they have? Who can tell by their definition? (They couldn't.) How many named gods were there? Again, "it all depends" on what you mean by "named". Whether the figure of 30,000 came from a lost manuscript of Varro, was a misinterpretation of the quoted manuscript, or from some other source, it would certainly have seemed to Augustine as perfectly reasonable. (Note that a much smaller number could also have been defended and would also have seemed reasonable to Augustine and his contemporaries.)