The Uncomfortable Truth

You've probably used a blockchain application recently. Swapped tokens on Uniswap, checked your NFT portfolio, tracked your DeFi positions. Felt good knowing you weren't depending on some institution to custody your assets or approve your transactions.

Here's what nobody tells you: you were trusting an institution the entire time.

Every time you interact with a blockchain app, you're not actually reading from the blockchain. You're reading from a middleman. Alchemy, Infura, The Graph's hosted service. These companies sit between you and the chain, and you're just trusting them to tell you the truth about what's happening onchain.

Your "trustless" trading bot? Trusts Alchemy's API. Your self-custody wallet? Trusts Infura to show you real balances. Your DAO governance dashboard? Trusts The Graph's servers aren't lying about votes.

We built cryptographic systems that eliminate trust in financial transactions, then immediately reintroduced trust requirements in how we read data about those transactions. It's embarrassing when you think about it.

Why Blockchains Suck at Answering Questions

Blockchains are optimized for consensus and immutability, not for queries. This isn't a bug. It's why they work.

Want your token balance? The blockchain doesn't store "balances." It stores every transaction that ever happened. Getting your balance means processing potentially millions of transactions, filtering for yours, doing the math, arriving at a number.

Top holders of an NFT collection? Scan every mint, every transfer, every burn in the contract's history. Then aggregate and rank.

A user's transaction history with USD values and decoded contract calls? Query the entire chain, identify relevant transactions, decode interactions, fetch metadata, calculate historical prices, format it all nicely.

Blockchains give you permissionless write access and cryptographically verifiable state transitions. But reading that state efficiently requires transformation work. Someone has to process raw blockchain data into something queryable. That someone is Alchemy, Infura, QuickNode, and a handful of other centralized companies.

And you're just trusting them. No verification. Just vibes.

What You're Actually Trusting

When you query an indexing service, you're trusting them in ways that directly contradict what blockchain technology is supposed to provide.

You're trusting they're correctly processing blockchain data. No way to verify cryptographically. You just believe them.

You're trusting they'll stay online. Third-party monitors have tracked over 1,000 Infura service incidents since 2022. That's roughly 250 per year. Five per week on average. Each time, the blockchains kept producing blocks. Consensus was fine. But applications couldn't read data without their centralized middleman. How can anyone call that reliable infrastructure for "trustless" applications?

You're trusting they won't restrict your access. But they can and do. API rate limits, authentication requirements, payment tiers. All forms of permissioned access to data that lives on permissionless blockchains.

You're trusting they're indexing what you need. But they decide what gets indexed and how. If certain queries are too expensive, they just don't offer them. Your application's capabilities are constrained by their business decisions.

You're trusting they won't censor. But they can filter or block whatever they want. If certain addresses or transactions don't appear in API responses, you'd never know.

And you're paying through the nose for all this trust. Projects burn hundreds of thousands or millions annually on indexing. It's like paying Instacart to buy your groceries, except you're not trading money for convenience. You're trading money for uncertainty. You're trusting they'll select the right items, that your avocado is actually ripe, that they didn't substitute something you didn't want. Except with blockchain data, they might be bringing you data that's just wrong, and you'd have no way to know until it's too late.

These costs scale with usage. The more successful you are, the more you pay. And unlike Instacart where you could just go to the store yourself, most developers don't have the option to "just go to the chain" directly without building expensive infrastructure. Some projects have shut down features or entire products because indexing costs became unsustainable.

The thing that should bother everyone: we've recreated exactly what blockchain technology was designed to eliminate. The need to trust intermediaries.

Why Existing "Solutions" Don't Actually Solve It

Let's be real about the options.

Centralized indexers like Alchemy and Infura are at least honest about what they are. Companies running infrastructure you pay to access. The developer experience is polished, but you're fully trusting them to represent blockchain state accurately, stay online, and not change their terms. You need API keys and payment. They can rate limit you, block you, go offline. There's no cryptographic verification. Just a business relationship. It's Web2 infrastructure for blockchain applications, which is fine if you're honest about it, but let's not pretend it's anything else.

"Decentralized" indexing networks like The Graph, SubQuery, and others don't actually solve the trust problem. They just distribute it. You still select specific indexers. You still trust they're serving accurate data. Token staking mechanisms incentivize honesty, but there's no cryptographic proof of correctness. If your chosen indexer serves bad data, you have no way to verify it without trusting another indexer. This isn't trustlessness. It's distributed trust. You're making explicit trust decisions about which entities provide your data, just spread across multiple nodes you picked yourself. The architecture still has single points of failure based on your selections. You're trusting economic incentives rather than using cryptographic proofs. It's decentralization theater.

Running your own infrastructure sounds good until you actually try it. Now the trust shifts to you, and you need to become a blockchain indexing expert whether you want to or not. You're no longer building your product. You're maintaining servers, optimizing database queries, debugging sync issues, staying current with node software updates. Every hour on indexing infrastructure is an hour not spent building features that differentiate your app and delight users.

Want to support multiple chains? Good luck. Ethereum has one set of challenges. Solana's account model is completely different. Bitcoin's UTXO model requires entirely different indexing approaches. Each chain has unique RPC quirks, state management patterns, performance characteristics you need to master. Your team of application developers becomes a team maintaining blockchain infrastructure. Your runway gets consumed by DevOps instead of product development. Your competitive advantage evaporates because you're solving the same indexing problems everyone else is solving instead of focusing on what makes your application unique.

You've become your own centralized indexer, with all the operational burden and none of the economies of scale. Worst of both worlds. You're still trusting someone (yourself and your team), but now you're paying full cost and opportunity cost.

Light clients provide verification but can't handle the queries modern applications need. They can verify data existed in a specific block through Merkle proofs, but they can't efficiently verify complex query results without processing significant blockchain history themselves. Not practical for finding all NFTs owned by an address, calculating portfolio values, tracking DeFi positions across protocols.

Archive nodes give you complete blockchain history but don't solve the query problem. You still need to process raw blockchain data into indexed form that applications can query efficiently. Archive nodes provide trustless access to raw transactions, not to the transformed, queryable data layer applications actually need.

None of these approaches deliver what blockchain technology is supposed to provide. They're workarounds making different trade-offs, not solutions that actually solve the problem.

The Contradiction We've Accepted

Let's be clear about what's happening architecturally. Blockchain validators are already running full nodes with complete blockchain state. They're processing every single transaction, executing smart contracts and generating events, maintaining the canonical, verifiable source of truth.

Then, separately, centralized companies run their own full nodes. They re-process all the same transactions. They re-extract all the same events. They re-structure the data into their own databases. They expose it through proprietary APIs that you just have to trust.

We're doing the work twice. Once in a trustless, verifiable way that produces the source of truth, and again in a trust-based way that produces the accessible version. We've taken data that comes with cryptographic guarantees and stripped those guarantees away in exchange for convenience.

Blockchain technology has matured. Applications are sophisticated, UX is improving, adoption is growing. But we're building on a foundation that fundamentally contradicts the technology's core value.

Every time someone uses a "trustless" application that depends on trusting Alchemy or Infura, they're experiencing the gap between what blockchain promises and what it delivers.

Smart contracts eliminate trust in financial transactions, but you trust API providers to know what's in those contracts. Private keys give you sovereign control of assets, but you need permission and payment to see what those assets are doing.

Two futures:

In one, blockchain infrastructure is trustless, permissionless, and verifiable all the way down. Applications cryptographically verify data without trusting providers. Users query blockchain state without permission. Censorship resistance extends from transaction processing to data access.

In the other, we've permanently compromised on trust at the most critical layer. "Trustless applications" remains marketing rather than technical reality.

The technology to fix this exists. The architecture works. The cryptographic foundations are solid. The question is whether we build it before the current model becomes so entrenched that escaping becomes impossible.

Shinzo is building the first future. We've gone back to first principles to design infrastructure where the read layer has the same properties as the write layer: trustless, permissionless, verifiable, sovereign.

In The Blockchain Middleman Problem: Fixing the Read Layer, we break down exactly how this works: the cryptographic foundations, the validator architecture, and the technical stack that makes verification replace trust.

Join the community building the data read layer blockchains deserve. X · Telegram · Discord · GitHub ·  shinzo.network