The simple answer is: this is not legal and also doesn't work at scale. Try running this type of scaling for a few thousand profiles - you will quickly be restricted.
It's definitely a breach of ToS, but I wouldn't be so fast at calling it illegal. It's a grey area that has yet to be properly litigated - I think the closest we've got is the LinkedIn scraping case and I don't remember whether that one even reached a conclusive answer.
In fact this is one of the downsides of the US legal system - litigation is so expensive that nobody dares trying it even though it could set a legal precedent that would benefit society at large. This is IMO something a consumer-friendly regulatory environment (such as the EU) should settle in advance like with the GDPR for example, but given they're not even bothered to enforce that effectively, I don't have much hope (if they enforced it, it would actually remove a big use-case for scraping Instagram, as you would be able to use the official clients without compromising your privacy).
You are wrong. This is not illegal. With an 4g/LTE proxy machine you can easily generate thousands of profiles rapidly and cheaply. They would be able to detect them at some point (will be harder if goes slowly) but it wouldn't stop the scraping.
The only way is for Instagram to restrict registration altogether, but you might create a black market where existing users sell their accounts, and cannabilize its own userbase (Bad for meta stock prices).
I may be wrong about this being illegal (depending on the country you reside in), but it is certainly not an approach that scales. Meta/Instagram have multiple teams dedicated to preventing this type of scraping. Unless you're willing to invest an equivalent level of resources, any success in scraping Instagram data will be temporary.
If there's demand for their service I don't see why it wouldn't scale. Get more phones, more SIM cards, and have automation around all this infrastructure to automate away as much of the stuff as possible.
> Meta/Instagram have multiple teams dedicated to preventing this type of scraping
That's great but ultimately they still have a weakness: they want people to be able to see their stuff - at least some of it - without logging in. As long as you can either simulate a normal device perfectly, or even better, use real devices or virtualize them, there isn't much that Facebook can do without impacting legitimate usage which they don't want.
