As of November 2025, there is not enough evidence to decisively judge this long‑horizon prediction about “the next several years and decades.”

On one hand, traditional centralized cloud and web platforms remain overwhelmingly dominant in usage and revenue. Public cloud infrastructure is still controlled primarily by a few hyperscalers: AWS, Microsoft Azure, and Google Cloud together hold roughly 60%+ of the global cloud infrastructure market, with AWS alone around 29–31% in 2024–2025 and the overall cloud market continuing to grow rapidly. (crn.com) Decentralized storage and compute networks (IPFS/Filecoin, Arweave, etc.) are tiny by comparison, and most consumer services (social, productivity, commerce, media, AI) still run on centralized architectures.

Decentralized and user‑controlled data models have grown, but from a small base and remain niche relative to mainstream web services. Web3 dapps saw strong growth, reaching on the order of 20–25 million daily unique active wallets by late 2024/early 2025, and hundreds of millions of people globally now hold crypto or use Web3 tools. Still, monthly active dapp users are estimated in the single‑digit millions to low tens of millions, far below the billions using centralized apps. (dappradar.com) Similarly, decentralized social protocols (Farcaster, Lens, Nostr) and federated networks (Mastodon, Bluesky) have grown to millions or at most tens of millions of users and a few million daily actives, which is meaningful but small compared to centralized platforms like X/Twitter, Facebook, Instagram, or TikTok. (blockeden.xyz) These numbers do not yet constitute decentralized services “winning” substantial market share versus centralized incumbents.

Moreover, much of what is marketed as “decentralized” today still depends heavily on centralized infrastructure. Many prominent Web3 applications and wallets rely on centralized RPC/node providers such as Infura and Alchemy, and on conventional cloud hosting (AWS, Google Cloud, Cloudflare) for front‑ends and APIs, creating de‑facto chokepoints. (forbes.com) That undercuts the idea that user‑controlled architectures have already structurally displaced enterprise‑controlled servers.

At the same time, there are visible trends in the direction Friedberg described. Major consumer platforms, especially Apple and Google, are increasingly emphasizing on‑device processing and privacy‑preserving designs—Apple Intelligence, for example, explicitly treats on‑device computation as its “cornerstone,” only escalating to Apple‑run servers when needed, while Google similarly pushes more on‑device AI and data minimization. (apple.com) Crypto wallets, non‑custodial identity, and decentralized apps have also grown substantially in users and technical maturity since 2022, even if they remain far from mainstream dominance. (dappradar.com) These developments support the direction of the prediction (more user‑centric architectures, more distributed control), but not yet its stronger claim that such models will win significant market share over centralized services.

Because Friedberg explicitly framed the timeline as “over the next couple of years and decades,” the horizon clearly extends well beyond 2025. The evidence so far shows modest but real growth of decentralized and user‑controlled models alongside continued dominance and innovation in centralized cloud and platform services. That mix is not enough to declare the prediction either clearly right or clearly wrong at this stage, so the fairest assessment is that it remains inconclusive (too early to tell).