SQD Network

SQD Network is a decentralized data infrastructure layer built to solve one of Web3's most persistent operational problems: getting reliable, complete, and verifiable blockchain data at scale without depending on a single centralized provider.

Developed by Subsquid Labs GmbH, a Swiss company incorporated in Zug and founded in 2021, SQD exposes its infrastructure through the Portal API — a plain HTTP interface that streams historical and live data from 200+ blockchain networks, including Solana, Ethereum, Base, Arbitrum, Bitcoin, and Hyperliquid. Behind Portal sits the SQD Network itself: a peer-to-peer storage and compute grid of 2,000+ worker nodes collectively holding more than two petabytes of indexed blockchain history at a 30x redundancy factor across the fleet.

How It Works

The network separates data ingestion from data serving. Subsquid Labs (the primary data provider in the current bootstrapping phase) pulls raw blocks from multiple independent RPC endpoints, runs cryptographic checks on every block, and compresses validated data into chunks. Those chunks are distributed by a central scheduler across worker nodes, each of which stores data locally and executes queries using DuckDB, a lightweight in-process analytical database.

When a consumer queries Portal, the request is routed to whichever workers hold the relevant data. Workers return results signed with their private keys, giving consumers a cryptographically verifiable audit trail for every response. The network targets under-50ms P90 latency and claims throughput 30x faster than querying historical data through standard RPC endpoints.

Validation Architecture

What differentiates SQD from conventional node providers is its rigorous treatment of data integrity. Rather than accepting whatever an RPC returns as ground truth, the ingester actively reconstructs and cross-checks each block:

• Transaction and receipt trie roots: Subsquid re-derives the Merkle Patricia Trie root from ingested transactions and compares it against the block header. Any missing or altered transaction changes the root and triggers a mismatch.
• Logs bloom filter: The 2,048-bit bloom filter is recomputed from ingested logs and compared byte-for-byte. Spurious or missing logs produce detectable divergence.
• Log index continuity: Sequential log indices (starting at 0) within a block must be unbroken; any gap indicates a missing log.
• Parent-hash chaining: Each block header commits to its predecessor, allowing verification of chain continuity from a single honest anchor.
• Multi-RPC arbitration: Disagreements between endpoints are resolved explicitly rather than silently accepting the first response.

A secondary system called Snoopy (currently in beta) adds post-ingestion fraud detection. It monitors whether different worker nodes return consistent results for identical queries — deterministic DuckDB outputs on the same Parquet data should never diverge. When they do, Snoopy generates cryptographic fraud proofs using SP1 zkVM-produced Groth16 proofs and submits them onchain to trigger slashing of the offending worker's bonded stake. This enforcement mechanism operates permissionlessly and does not require SQD's team to intervene.

Solana Integration

Solana is a first-class supported chain within Portal and the broader SQD developer toolkit. The Portal API exposes a Solana-specific endpoint covering program instructions, token transfers, transaction-level data, and account balances — making it possible to query token transfers on specific program IDs (such as the SPL Token program) with plain HTTP requests and no setup beyond an API key. The Solana Squid SDK provides a structured indexing framework for building custom processors over Solana data — supporting transaction, instruction, balance, and reward data streams — along with a typegen tool for type-safe instruction decoding. The Pipes SDK extends this to a streaming model for building real-time Solana data pipelines without managing validator or RPC infrastructure directly.

Products and Developer Tooling

Beyond Portal, SQD operates several adjacent products consolidated under the sqd.dev domain in May 2026:

• SQD Cloud: A managed deployment environment for running Squid indexers, with monitoring through the SQD 360 dashboard.
• Squid SDK: An open-source multi-chain indexing framework supporting EVM, Solana, Substrate, and Tron.
• Pipes SDK: A lower-level streaming interface for building custom data pipelines directly from indexed blockchain data.
• Portal MCP Server: An integration enabling AI agents to query onchain data via the Model Context Protocol, positioning SQD as infrastructure for autonomous agent workflows.

SQD Token

The SQD token serves four distinct functions within the network:

Worker bonding: Operators must stake 100,000 SQD to run a worker node. This bond is slashable — Snoopy fraud proofs can trigger automated on-chain slashing of misbehaving workers.

Delegation: Token holders can delegate to trusted worker operators, sharing in rewards proportional to delegated stake. The protocol whitepaper targets approximately 10% annual yield for delegators and 20–30% for operator-workers depending on bonded capital.

Rate limiting and bandwidth allocation: Gateway operators lock tokens to receive increased query throughput measured in compute units. Longer lock periods apply a booster multiplier to bandwidth allocation — for example, a 3-year lock yields a significantly higher base query rate per token than a shorter commitment.

Revenue Pools: This mechanism ties SQD token demand directly to real data usage. Enterprise clients pay for data access in stablecoins. Incoming fees are automatically split by a FeeRouter smart contract: 50% remains as stablecoin rewards for liquidity providers, 45% is converted to SQD via automated swaps on PancakeSwap V3 and routed to worker rewards, and 5% is converted to SQD and permanently burned. The automated swap uses a 30-minute TWAP oracle with a 3% slippage ceiling to protect against price manipulation. This creates ongoing token buy pressure from real data demand rather than speculation.

Team and Clients

Following a June 2026 leadership transition, SQD operates under CEO Wanja S. Oberhof, who brings a capital markets background, alongside Head of Transformation Robert Jung, a former EY partner. Founder Marcel Fohrmann continues to lead protocol and ecosystem development. Co-founder and CTO Dr. Dimitrii Zhelezov serves on the board alongside Jordi Alexander (founder of Selini Capital), Davey Airey (Chairman), and board members from capital markets and AI backgrounds.

As of mid-2026, the company reported processing 1.4 petabytes of data per day across 225+ networks, serving protocols managing over $11 billion in total value locked. Named clients include Morpho, GMX, PancakeSwap, RAILGUN, and Deutsche Telekom.

Ecosystem Fit

For the Solana ecosystem, SQD addresses a genuine infrastructure gap. Building real-time or historical analytics, indexers, or data pipelines on Solana typically requires either managing your own validator or RPC infrastructure, or depending on centralized providers whose data correctness cannot be independently verified. SQD's worker-signed responses and multi-layer cryptographic validation offer an alternative with a verifiable audit trail — one that extends to historical queries at depths that most standard RPC providers do not retain. The Squid SDK and Pipes SDK lower the barrier for developers who need structured access to Solana program instruction and token transfer data without standing up indexing infrastructure themselves, while the Portal MCP Server opens the same data layer to AI agents that need on-demand access to Solana blockchain history.