Instagram weights session continuity heavily. A request arriving from a network the account has never used, paired with a browser fingerprint it has never seen, is the combination most likely to produce a login challenge. Changing the IP mid-session is more disruptive here than on most platforms, which makes rotation the wrong default.
Its public surface is also narrower than it used to be. Profile and hashtag pages that once loaded anonymously increasingly require an authenticated session, so a large amount of what people describe as Instagram scraping now means authenticated access, and that is governed by Instagram's terms rather than by your proxy configuration.
For data about assets you own, the Instagram Graph API is the supported route and needs no proxy at all. Proxies belong to the workflows the API does not cover: checking how a public profile or ad renders from a specific market, and verifying regional delivery.
It is worth being clear about when a proxy is the wrong tool here. If you are managing accounts you own, Instagram's own tooling and the Graph API cover it without a proxy. If you are trying to operate accounts at a scale the platform does not permit, no network configuration makes that durable, and the cost of a restriction usually exceeds whatever the workflow was worth. Proxies earn their place in observation work: seeing what a market sees, verifying delivery, and checking public rendering from outside your own network.
If you're comparing workflows before you commit, explore all proxy use cases to see which setup fits best.