I have a question about the statement of Remark 1.4 following on Thm 1.3 from Burt Totaro's paper "The Chow Ring of a Classifying Space" (p. 4):
Let $G$ be a reductive group over a field $k$.
Remark 1.4. If $G$ is finite, then the geometric quotient $V/G$ exists as an affine variety for all representations $V$ of $G$ [28: Mumford's GIT]. So $(V -S)/G$ is a quasi-projective variety for all closed subsets $S \subset V$ such that $G$ acts freely on $V - S$.
Question: Does anybody see to which explicit result from GIT Totaro refers to in his remark that for every closed subsets $S \subset V$ such that $G$ acts freely on $V - S$, the quotient $(V -S)/G$ exist as a quasi-projective variety? Clearly that's the same as to consider all open subsets $U \in V$ (which can be always canonically endowed with structure of an variety via open immersion; see https://stacks.math.columbia.edu/tag/01IM) with free $G$-action.
Nevertheless in GIT book I haven't found a result with these requirements implying quasi-projectivity of $U/G$. Does anybody know which GIT result Totaro has here in mind?