It's the pet theory of several French classicists (primarily Christian Goudineau), partly out of a desire to reverse the hero-worship that was built around Vercingetorix in the nationalist atmosphere of post-1871 France, where he became through some circuitous reasoning a symbol of the "revanchard" spirit against the Prussians.

In actuality, there's very little in the contemporary sources that suggests it was the case (Caesar makes no mention of it, and neither does Strabo). Goudineau's case is based on a much later account by Cassius Dio in his History of Rome, who states that Vercingetorix hoped that Caesar would show mercy because of their former friendship, and later that "Caesar reproached his treason to the surrendering Vercingetorix, opposing his [Vercingetorix's] rebellion to his own friendship". Interpreting these passages as meaning that Vercingetorix and Caesar had a preexisting relationship isn't necessarily wrong, but they could also refer to a more abstract bond since the Arverni had made a treaty with Rome after their defeat in 121 BC. And of course, Cassius Dio wrote more than a century after the war, making a fine interpretation of the wording of a couple sentences in his work a bit of a zero-sum game.

It's not impossible that Vercingetorix served as an auxiliary in his youth (Romans using Gallic cavalry is well-evidenced), and some historians see his grasp of Roman tactics as compelling evidence for it, but there really isn't much of a proof of it.

e: Going back to this to correct a slightly wonky mental calculation on my part: Cassius Dio's History of Rome was written more than two centuries after the Gallic Wars, not one.