Summary of the prediction Jason predicted that growing worries about government access to cloud‑hosted data would cause “a lot of people” to move off the cloud, with (a) many consumers and small businesses buying/using self‑hosted “mini servers” or similar local encrypted storage, and (b) broader adoption of cloud/device services where providers don’t hold the decryption keys.

What actually happened (2022–2025)

1. Cloud usage kept growing strongly instead of people moving off it.

   • Worldwide IaaS/public‑cloud infrastructure revenue rose from about $120B in 2022 to ~$140B in 2023 and ~$172B in 2024, with double‑digit annual growth, and analysts expect continued >20% growth. (gartner.com)
   • The overall cloud storage market itself was about $59B in 2024 and projected to exceed $270B by 2031. (globenewswire.com)
     This is the opposite of “a lot of people moving off of the cloud”; mainstream behavior has been more cloud, not less.

2. Self‑hosted "mini servers"/NAS did grow, but remain a relatively small niche.

   • Consumer/home NAS is a growing category, with global consumer NAS revenue estimated around $6B in 2024 and projected to roughly double by 2030; residential use is the fastest‑growing segment, and privacy plus centralised home storage are cited as drivers. (grandviewresearch.com)
   • By contrast, cloud storage services are tens of billions of dollars already and growing faster in absolute terms than consumer NAS, meaning NAS remains a comparatively small slice of how people store data. (globenewswire.com)
     So while there is a visible home‑lab / private‑cloud niche, it does not amount to a broad consumer or small‑business migration away from mainstream cloud providers.

3. End‑to‑end encrypted / “we don’t have the keys” services expanded, but mostly as options layered on existing clouds.

   • Apple introduced Advanced Data Protection for iCloud in late 2022 and expanded it globally in 2023, letting users keep keys on their own devices so Apple cannot access most iCloud categories—if they opt in. Apple’s own documentation stresses it’s an optional setting, not the default. (macrumors.com)
   • Meta began rolling out default end‑to‑end encryption for Messenger in 2023, after years of offering E2EE only as an opt‑in, and plans to extend this to Instagram messaging. (techcrunch.com)
   • Privacy‑first providers like Proton (Mail/Drive/Docs/etc.), which are architected so the provider cannot read user content, grew to over 100 million accounts by 2023. (en.wikipedia.org)
     These moves do validate the direction Jason described for part (b), but they remain optional features or minority platforms layered on top of a still‑booming conventional cloud ecosystem. There is no evidence that the typical consumer or small business has shifted to “provider‑can’t‑decrypt” architectures for most of their data.

4. Government access pressures did increase, but didn’t trigger a broad behavioral shift.

   • Transparency and independent analyses show sharp growth in government data requests: a Proton analysis of Apple, Google, and Meta transparency reports found accounts handed to US authorities rose over 600% from 2014–2024, fueling concern about state access to cloud‑stored data. (euronews.com)
   • The UK’s use of the Investigatory Powers Act to secretly demand backdoor access to Apple’s end‑to‑end encrypted iCloud backups led Apple to pull its Advanced Data Protection feature in the UK rather than weaken encryption—an emblematic clash over government access. (theverge.com)
     Despite these headline cases, the market data above shows continued rapid expansion of mainstream cloud services and only niche‑scale growth of self‑hosted alternatives. So the political pressure and concern Jason cited are real, but they have not yet translated into a mass exodus from the cloud.

Bottom line By late 2025, we do see:

• More government attempts to get at cloud data, and
• Some growth in self‑hosted storage and in services where providers don’t hold decryption keys.

However, these shifts are small compared to the explosive overall growth of conventional cloud usage, and most consumers and small businesses still rely on major cloud providers that can access their data. That means the core prediction—"a lot of people moving off of the cloud" toward mini‑servers and fully user‑keyed services—has not materialized at scale.

Verdict: the prediction is mostly wrong.