“The internet was designed to be open, but the platforms built on top of it are not.”
——Chris Dixon, Rebooting the Internet
0. From Open Web1 to Centralized Web2: The Legacy of Missing Trust
The Web1 era began with openness. Born out of academic and military collaboration, the TCP/IP protocol stack laid the foundation for global connectivity. TCP/IP was — and remains — an open and permissionless stack: any device following the protocol can join the network. This property of permissionless connectivity created the early decentralized flavor of the Internet.
But as the Internet commercialized, new demands emerged: identity, content, value transfer. We soon realized that TCP/IP solved the problem of connectivity, but not the problem of trust.
For example:
- How do users authenticate their identities?
- Who owns or controls the data?
- How can cross-site coordination be trusted?
- How do we prevent double-spending or denial in financial interactions?
Because trust was never embedded into the network stack, applications were forced to implement it themselves. This gap was filled by centralized platforms. Thus, Web2 emerged — an architecture where platforms manage identities, content, and transaction integrity, becoming central custodians of user data and trust.
1. TCP/IP: The Backbone of the Open Internet
TCP/IP is the foundational communication protocol suite of the Internet. It is organized as a layered architecture:
| TCP/IP Layer | OSI Equivalent | Key Protocols | Role |
|---|---|---|---|
| Application | Application | HTTP, DNS | High-level application logic |
| Transport | Transport | TCP, UDP | End-to-end transmission |
| Network | Network | IP, ICMP | Routing and addressing |
| Link | Data/Physical | Ethernet, Wi-Fi | Physical transmission |
The stack’s key advantage is abstraction and decoupling: each layer only depends on the layer below. This modularity enabled the Internet’s rapid expansion.
However, the TCP/IP philosophy is based on best-effort delivery:
The lower layers do not guarantee delivery; higher layers are responsible for reliability.
For example, the IP layer may drop or reorder packets; TCP adds retransmission and reordering to give the illusion of reliability. In Web2, this is acceptable because platforms have centralized control. But Web3 does not have that luxury.
2. Web3 Demands More Than “Best Effort” — It Demands Cryptographic Guarantees
Web3 aims to build not an Internet of information, but an Internet of value. That means communication isn’t just about moving bytes, but about moving state, value, commitments — all in a verifiable, tamper-resistant, trust-minimized way.
Why TCP/IP Falls Short for Web3
| Core Web3 Requirement | Can TCP/IP Provide It? |
|---|---|
| Verifiable state transitions | ❌ Stateless |
| Decentralized identity | ❌ No identity layer |
| Cross-chain communication | ❌ No context awareness |
| Zero-knowledge proof transport | ❌ No crypto semantics |
| Permissionless composability | ❌ No execution model |
| Censorship-resistant P2P messaging | ❌ No built-in resistance |
In short, TCP/IP may be the physical substrate, but it is not Web3-native. Today, much of Web3 still relies on centralized infrastructure for critical operations:
- Wallets interact with chains via centralized RPC providers (e.g., Infura)
- IPFS content often gets routed through centralized gateways
- Smart contract calls are initiated by centralized frontends
Web3 today is still a Web3 emulator running on Web2 protocols.
3. A Comparative Stack: Web2 vs Web3 Protocol Layers
To reframe Web3’s architecture, we can draw an analogy with TCP/IP:
| Information Internet | Protocol Layer | Value Internet (Web3) | Protocol Layer |
|---|---|---|---|
| IP addresses | IP (Network Layer) | Blockchain accounts | Layer 1 state machines |
| Packet transport | TCP / UDP (Transport) | Rollups / Channels | Layer 2 state extensions |
| Website access | HTTP / DNS (App Layer) | DApps, NFTs, DAOs | Application protocols (ERC-20/721) |
What’s missing from Web3 is not just more dApps — it’s a native backbone layer, a “TCP/IP-equivalent” for trust, identity, and value. We need a Layer 0 purpose-built for Web3.
4. Why Didn’t TCP/IP Include This?
When TCP/IP was developed in the 1970s, it was designed to solve the problem of reliable connectivity, not trust. Many technologies needed for decentralized trust simply didn’t exist yet:
| Capability | Missing in TCP/IP Era | Web2 Replacement |
|---|---|---|
| Decentralized identity | No public-key addresses / DID | Email + password (OAuth) |
| Decentralized storage | No IPFS / Filecoin | AWS / GCP |
| Decentralized compute | No zkVM / decentralized VMs | Serverless / Cloud |
| Decentralized payments | No cryptocurrencies / smart contracts | Credit card / PayPal |
So although TCP/IP was open and permissionless, the application layer had no choice but to rely on centralized platforms to provide closed-loop services. The result was Web2 — an ecosystem where:
- Users generate data
- Platforms own data
- Algorithms control users
This model has been called data feudalism — users become digital serfs working the land owned by platforms.
5. Web3 Is a Return to TCP/IP’s Spirit — Not a Rejection
Web3 is not trying to overthrow the Internet — it’s trying to complete what TCP/IP never finished: embedding trust into the stack.
We can envision a new decentralized protocol stack:
| Layer | Web2 Implementation | Web3 Equivalent |
|---|---|---|
| Communication | TCP/IP | TCP/IP (shared) |
| Structure | JSON / SQL | Merkle Trees / SMTs |
| Identity | OAuth (Google, Facebook) | Public-key addresses / DIDs |
| Storage | AWS / Cloud | IPFS / Arweave / Filecoin |
| Compute | Cloud Functions | EVM / zkVM / WASM + ZK |
| Consensus | Centralized databases | Blockchain (PoW / PoS / BFT) |
| Application | Web2 platforms | DApps / DAOs / DeFi / NFTs |
Web3’s mission is to use cryptography + distributed systems to build native, verifiable protocols for trust and computation.
6. So What Should Web3’s TCP/IP Look Like?
We need a Web3-native Layer 0 stack that transmits not just packets, but:
- Verifiable state snapshots
- Censorship-resistant messages
- Zero-knowledge proofs and public-key identities
- Deployable, composable contracts
- Cross-chain state interoperability
Projects like Celestia, Eigenlayer, and IBC are exploring this territory — building the “TCP/IP for the Internet of Value”.
7. Web3’s “New TCP/IP” Isn’t a Rebellion — It’s a Foundation
Web1 gave us protocol freedom
Web2 gave us platform convenience
Web3 seeks to reclaim the protocol layer — without relying on the platform
Web3 doesn’t aim to replace TCP/IP — it aims to augment it with a native, trust-aware state layer.
This new “connectivity” layer won’t just route data packets — it will route state, trust, value, and cryptographic commitments. We are no longer content with “best-effort”. What we demand is proof-driven delivery.
And that is the foundational infrastructure Web3 is still missing.