That depends on what you mean be Google-free, and what you mean by sustainable.
If you mean Google-free in terms of minimizing interactions with Google’s servers and proprietary code, then somewhat paradoxically, it seems like Google’s Pixel phones are the de facto gold standard for deGoogleable phones, perhaps because Google is most at risk of accusations of anti-competitive practices on its own phones. CalyxOS seems primarily designed for Pixels, with the FP4 as a second option (they warn that security updates for propriety components are often outdated). GrapheneOS only supports Pixels. Almost every independent Android distribution supports them, and they seem to have the best support for alternative Android distributions.
If you mean Google-free in terms of no interactions with Google servers, then you start having to worry about all the details discussed in the comments above. Often there are details, and remaining interactions, that would be significant detriments to usability to remove, and so they end up staying in. Actually avoiding all interactions with Google is extremely difficult: for example, as ReCaptcha involves agreeing with Google terms and interacting with Google, there are many state services, including paying some taxes, that I can’t easily do through official websites without also interacting with Google. At this level, you really have to consider the tradeoffs involved in avoiding each of these interactions. And apart from these interactions, proprietary firmware blobs show up everywhere.
If you mean Google-free in terms of no significant Google code at all, then Android as a whole, being a Google product, isn’t suitable. As much as it is nominally open, it is controlled by Google, not community-led, and is deeply integrated with Google: alternative Android distributions must work to remove those integrations without making the system extremely impractical, with Google as an adversary putting up barriers to doing so. You’d likely need to look at non-Android OSs, like postmarketOS, Ubuntu Touch, PureOS, Plasma Mobile, and others, and non-Android phones, like the PinePhone or Purism Librem 5. These are generally significantly harder to use and less stable than Android phones; even basic functionality can sometimes be problematic. On the other hand, when they do work, they both offer far better control of interactions with Google and other companies, and can often run potentially problematic Android apps in sandboxes to help avoid unwanted behaviours.
I think a reasonable summation of Fairphone, for that sort of sustainability focused on useful longevity and long-term repairability, is that their phones are significantly better than most other phones, and significantly worse than their marketing suggests.
But in general, if you’re interested in long-term repairability, first-party support and parts availability does usually end, eventually, even if the timescales involved are usually far shorter for modern consumer electronics than other products*. Fairphone does usually provide first-party parts for longer than other manufacturers.
After that point, repairs become matters of doing more component-level repairs with available or constructed components, finding third-party-manufactured or old stock parts, or taking parts from other, usually broken, devices.
For the latter two, somewhat unfortunately, somewhat regardless of the original support offered, more popular products from bigger manufacturers often end up being more repairable eventually, because more parts were originally available, more devices being made means more are likely available to use as sources of parts, more popularity often means more independent reverse engineering and information, and more third-party manufacturers may have made parts as well. As a longer-term example, while Waltham and Elgin are just as dead as many smaller watch manufacturers, even without the wide availability of parts from broken watches for Waltham and Elgin, you can rather easily find original spare parts in original packaging, and third-party parts and reproductions, while parts for smaller manufacturers may well need to be machined. There was simply much more made—both products and parts—and after well over a half-century of depletion, the stocks of the more highly produced products end up lasting longer.
Thus, I would suspect that you would now have an easier time repairing, say, a Nexus 4, or even, quite possibly, an iPhone from the earliest years, than a Fairphone 1. Over time, the factors involved simply change, unfortunately: even if one device is designed to be more repairable, if parts eventually become unavailable, the more repair-hostile device may become the easier one to repair. I recently bought a replacement battery for a phone from 2007, and a number of parts—of course, of very variable and often dubious quality—for the Nexus 4 and iPhone 4 are on AliExpress, while third-party parts even for the Fairphone 2 simply don’t seem to exist, much less the Fairphone 1.
(* While probably an extreme case, I once dealt with a product where we had support, and usually a support contract, from the manufacturer, for 120 years (for organizational reasons, we only have documentation back to the mid-1920s, but almost certainly had a support contract before then as well). Then they suddenly decided, a few years ago, that they didn’t want to support their oldest products any longer (I don’t know what the cutoff was, maybe 1900?), and unexpectedly wouldn’t renew our support contract. To give some context on what extremely-long-term support ends up looking like, the company providing support to us now stockpiles parts they scavenge throughout the country from devices being dismantled, and even then, many repairs involve them having replacement parts individually machined. For reasons of historical significance, we have absolutely no intention of ever replacing the device, and likely couldn’t even do so legally.)